Collaborative Filtering(CF) is a leading approach to build recommender systems which has gained considerable development and popularity. A predominant approach to CF is rating prediction recommender algorithm, aiming ...Collaborative Filtering(CF) is a leading approach to build recommender systems which has gained considerable development and popularity. A predominant approach to CF is rating prediction recommender algorithm, aiming to predict a user's rating for those items which were not rated yet by the user. However, with the increasing number of items and users, thedata is sparse.It is difficult to detectlatent closely relation among the items or users for predicting the user behaviors. In this paper,we enhance the rating prediction approach leading to substantial improvement of prediction accuracy by categorizing according to the genres of movies. Then the probabilities that users are interested in the genres are computed to integrate the prediction of each genre cluster. A novel probabilistic approach based on the sentiment analysis of the user reviews is also proposed to give intuitional explanations of why an item is recommended.To test the novel recommendation approach, a new corpus of user reviews on movies obtained from the Internet Movies Database(IMDB) has been generated. Experimental results show that the proposed framework is effective and achieves a better prediction performance.展开更多
This paper examines the prediction of film ratings.Firstly,in the data feature engineering,feature construction is performed based on the original features of the film dataset.Secondly,the clustering algorithm is util...This paper examines the prediction of film ratings.Firstly,in the data feature engineering,feature construction is performed based on the original features of the film dataset.Secondly,the clustering algorithm is utilized to remove singular film samples,and feature selections are carried out.When solving the problem that film samples of the target domain are unlabelled,it is impossible to train a model and address the inconsistency in the feature dimension for film samples from the source domain.Therefore,the domain adaptive transfer learning model combined with dimensionality reduction algorithms is adopted in this paper.At the same time,in order to reduce the prediction error of models,the stacking ensemble learning model for regression is also used.Finally,through comparative experiments,the effectiveness of the proposed method is verified,which proves to be better predicting film ratings in the target domain.展开更多
Fossil fuels are undoubtedly important, and drilling technology plays an important role in realizing fossil fuel exploration;therefore, the prediction and evaluation of drilling efficiency is a key research goal in th...Fossil fuels are undoubtedly important, and drilling technology plays an important role in realizing fossil fuel exploration;therefore, the prediction and evaluation of drilling efficiency is a key research goal in the industry. Limited by the unknown geological environment and complex operating procedures, the prediction and evaluation of drilling efficiency were very difficult before the introduction of machine learning algorithms. This review statistically analyses rate of penetration(ROP) prediction models established based on machine learning algorithms;establishes an overall framework including data collection, data preprocessing, model establishment, and accuracy evaluation;and compares the effectiveness of different algorithms in each link of the process. This review also compares the prediction accuracy of different machine learning models and traditional models commonly used in this field and demonstrates that machine learning models are the most effective technical means in current ROP prediction modeling.展开更多
Recommender systems are rapidly transforming the digital world into intelligent information hubs.The valuable context information associated with the users’prior transactions has played a vital role in determining th...Recommender systems are rapidly transforming the digital world into intelligent information hubs.The valuable context information associated with the users’prior transactions has played a vital role in determining the user preferences for items or rating prediction.It has been a hot research topic in collaborative filtering-based recommender systems for the last two decades.This paper presents a novel Context Based Rating Prediction(CBRP)model with a unique similarity scoring estimation method.The proposed algorithm computes a context score for each candidate user to construct a similarity pool for the given subject user-item pair and intuitively choose the highly influential users to forecast the item ratings.The context scoring strategy has an inherent capability to incorporate multiple conditional factors to filter down the most relevant recommendations.Compared with traditional similarity estimation methods,CBRP makes it possible for the full use of neighboring collaborators’choice on various conditions.We conduct experiments on three publicly available datasets to evaluate our proposed method with random user-item pairs and got considerable improvement in prediction accuracy over the standard evaluation measures.Also,we evaluate prediction accuracy for every user-item pair in the system and the results show that our proposed framework has outperformed existing methods.展开更多
Recently,internet users have significantly increased their use of search engines,and market investors are no exception.As a result,predictive models that incorporate scattered web-based information are developing as a...Recently,internet users have significantly increased their use of search engines,and market investors are no exception.As a result,predictive models that incorporate scattered web-based information are developing as an area of forecasting.The objective of this research is to compare the predictive accuracy of fundamental macroeconomic variables,online attention series measured by the Google Trends search volume index,and a combination of both data types for the Mexican,Brazilian,Chilean,and Colombian currencies paired with the USD.The exchange rate series used in this study are sourced from a real-time platform.Four indicators capturing the fundamental macroeconomic differences between these emerging economies and the U.S.from January 2004 to March 2021(monthly)were analyzed.To assess the predictive performance of the KNN algorithm,OLS regression and the random walk with drift model were compared.Considering in-sample predictions,the results generally exhibit lower estimation errors in the random walk with drift model,but in the joint fundamental–online attention data,the KNN and OLS predictions are more accurate than those of the random walk with drift.However,the KNN predictions based on out-of-sample fit generate the lowest estimation errors and the most accurate predictions for the joint fundamental–online attention data.Additionally,performance testing indicates that the KNN extended model outperforms the out-ofsample forecast for the OLS regression and the random walk with drift model.展开更多
The effectiveness of recommendation systems heavily relies on accurately predicting user ratings for items based on user preferences and item attributes derived from ratings and reviews.However,the increasing presence...The effectiveness of recommendation systems heavily relies on accurately predicting user ratings for items based on user preferences and item attributes derived from ratings and reviews.However,the increasing presence of fake user data in these ratings and reviews poses significant challenges,hindering feature extraction,diminishing rating prediction accuracy,and eroding user trust in the system.To tackle this issue,we propose a robust rating prediction model for recommendation systems that integrates fake user detection and multi-layer feature fusion.Our model utilizes a GraphSAGE-based submodel to filter out fake user data from rating data and review texts.To strengthen fake user detection,we enhance GraphSAGE by selecting aggregation neighbors based on the collusion fraud degree among users,and employ an attention mechanism to weigh the contribution of each neighbor during representation aggregation.Furthermore,we introduce a multi-layer feature fusion submodel to integrate diverse features extracted from the filtered ratings and reviews.For deep feature extraction from review texts,we implement a temporal attention mechanism to analyze the relevance of reviews over time.For shallow feature extraction from rating data,we incorporate trust evaluation mechanism and cloud model to assess the influence of trusted neighbors’ratings.In our evaluation,we compare our model against six baseline models for fake user detection and four rating prediction models across five datasets.The results demonstrate that our model exhibits significant performance advantages in both fake user detection and rating prediction.展开更多
Sensor selection and optimization is one of the important parts in design for testability. To address the problems that the traditional sensor optimization selection model does not take the requirements of prognostics...Sensor selection and optimization is one of the important parts in design for testability. To address the problems that the traditional sensor optimization selection model does not take the requirements of prognostics and health management especially fault prognostics for testability into account and does not consider the impacts of sensor actual attributes on fault detectability, a novel sensor optimization selection model is proposed. Firstly, a universal architecture for sensor selection and optimization is provided. Secondly, a new testability index named fault predictable rate is defined to describe fault prognostics requirements for testability. Thirdly, a sensor selection and optimization model for prognostics and health management is constructed, which takes sensor cost as objective function and the defined testability indexes as constraint conditions. Due to NP-hard property of the model, a generic algorithm is designed to obtain the optimal solution. At last, a case study is presented to demonstrate the sensor selection approach for a stable tracking servo platform. The application results and comparison analysis show the proposed model and algorithm are effective and feasible. This approach can be used to select sensors for prognostics and health management of any system.展开更多
Rate of penetration(ROP) of a tunnel boring machine(TBM) in a rock environment is generally a key parameter for the successful accomplishment of a tunneling project. The objectives of this work are to compare the accu...Rate of penetration(ROP) of a tunnel boring machine(TBM) in a rock environment is generally a key parameter for the successful accomplishment of a tunneling project. The objectives of this work are to compare the accuracy of prediction models employing partial least squares(PLS) regression and support vector machine(SVM) regression technique for modeling the penetration rate of TBM. To develop the proposed models, the database that is composed of intact rock properties including uniaxial compressive strength(UCS), Brazilian tensile strength(BTS), and peak slope index(PSI), and also rock mass properties including distance between planes of weakness(DPW) and the alpha angle(α) are input as dependent variables and the measured ROP is chosen as an independent variable. Two hundred sets of data are collected from Queens Water Tunnel and Karaj-Tehran water transfer tunnel TBM project. The accuracy of the prediction models is measured by the coefficient of determination(R2) and root mean squares error(RMSE) between predicted and observed yield employing 10-fold cross-validation schemes. The R2 and RMSE of prediction are 0.8183 and 0.1807 for SVMR method, and 0.9999 and 0.0011 for PLS method, respectively. Comparison between the values of statistical parameters reveals the superiority of the PLSR model over SVMR one.展开更多
In this paper, the method which can combine different seismic data with the different precision and completeness, even the palaeo-earthquake data, has been applied to estimate the yearly seismic moment rate in the sei...In this paper, the method which can combine different seismic data with the different precision and completeness, even the palaeo-earthquake data, has been applied to estimate the yearly seismic moment rate in the seismic region. Based on this, the predictable model of regional time-magnitude has been used in North China and Southwest China. The normal correlation between the time interval of the events and the magnitude of the last strong earthquake shows that the model is suitable. The value of the parameter c is less than the average value of 0.33 that is obtained from the events occurred in the plate boundary in the world. It is explained that the correlativity between the recurrence interval of the earthquake and the magnitude of the last strong event is not obvious. It is shown that the continental earthquakes in China are different from that occurred in the plate boundary and the recurrence model for the continental events are different from the one for the plate boundary events. Finally the seismic risk analysis based on this model for North China and Southwest China is given in this paper.展开更多
The reliability of electrical connectors has critical impact on electronic systems. It is usually characterized by failure rate prediction value according to standard MIL-HDBK-217(or GJB-299 C in Chinese) in engineeri...The reliability of electrical connectors has critical impact on electronic systems. It is usually characterized by failure rate prediction value according to standard MIL-HDBK-217(or GJB-299 C in Chinese) in engineering practice. Given to their limitations and mislead results, a new failure rate prediction models needs to be presented. The presented model aims at the mechanism of increase of film thickness which leads to the increase of contact resistance. The estimated failure rate value can be given at different environmental conditions,and some of the factors affecting the reliability are taken into account. Accelerated degradation test(ADT) was conducted on GJB599 III series electrical connector. The failure rate prediction model can be simply formed and convenient to calculate the expression of failure rate changing with time at various temperature and vibration conditions. This model gives an objective assessment in short time, which makes it convenient to be applied to the engineering.展开更多
In the era of intelligent economy, the click-through rate(CTR) prediction system can evaluate massive service information based on user historical information, and screen out the products that are most likely to be fa...In the era of intelligent economy, the click-through rate(CTR) prediction system can evaluate massive service information based on user historical information, and screen out the products that are most likely to be favored by users, thus realizing customized push of information and achieve the ultimate goal of improving economic benefits. Sequence modeling is one of the main research directions of CTR prediction models based on deep learning. The user's general interest hidden in the entire click history and the short-term interest hidden in the recent click behaviors have different influences on the CTR prediction results, which are highly important. In terms of capturing the user's general interest, existing models paid more attention to the relationships between item embedding vectors(point-level), while ignoring the relationships between elements in item embedding vectors(union-level). The Lambda layer-based Convolutional Sequence Embedding(LCSE) model proposed in this paper uses the Lambda layer to capture features from click history through weight distribution, and uses horizontal and vertical filters on this basis to learn the user's general preferences from union-level and point-level. In addition, we also incorporate the user's short-term preferences captured by the embedding-based convolutional model to further improve the prediction results. The AUC(Area Under Curve) values of the LCSE model on the datasets Electronic, Movie & TV and MovieLens are 0.870 7, 0.903 6 and 0.946 7, improving 0.45%, 0.36% and 0.07% over the Caser model, proving the effectiveness of our proposed model.展开更多
This research concentrates to model an efficient thyroid prediction approach,which is considered a baseline for significant problems faced by the women community.The major research problem is the lack of automated mod...This research concentrates to model an efficient thyroid prediction approach,which is considered a baseline for significant problems faced by the women community.The major research problem is the lack of automated model to attain earlier prediction.Some existing model fails to give better prediction accuracy.Here,a novel clinical decision support system is framed to make the proper decision during a time of complexity.Multiple stages are followed in the proposed framework,which plays a substantial role in thyroid prediction.These steps include i)data acquisition,ii)outlier prediction,and iii)multi-stage weight-based ensemble learning process(MS-WEL).The weighted analysis of the base classifier and other classifier models helps bridge the gap encountered in one single classifier model.Various classifiers aremerged to handle the issues identified in others and intend to enhance the prediction rate.The proposed model provides superior outcomes and gives good quality prediction rate.The simulation is done in the MATLAB 2020a environment and establishes a better trade-off than various existing approaches.The model gives a prediction accuracy of 97.28%accuracy compared to other models and shows a better trade than others.展开更多
In recent years,deep learning has been widely applied in the fields of recommendation systems and click-through rate(CTR)prediction,and thus recommendation models incorporating deep learning have emerged.In addition,t...In recent years,deep learning has been widely applied in the fields of recommendation systems and click-through rate(CTR)prediction,and thus recommendation models incorporating deep learning have emerged.In addition,the design and implementation of recommendation models using information related to user behavior sequences is an important direction of current research in recommendation systems,and models calculate the likelihood of users clicking on target items based on their behavior sequence information.In order to explore the relationship between features,this paper improves and optimizes on the basis of deep interest network(DIN)proposed by Ali’s team.Based on the user behavioral sequences information,the attentional factorization machine(AFM)is integrated to obtain richer and more accurate behavioral sequence information.In addition,this paper designs a new way of calculating attention weights,which uses the relationship between the cosine similarity of any two vectors and the absolute value of their modal length difference to measure their relevance degree.Thus,a novel deep learning CTR prediction mode is proposed,that is,the CTR prediction network based on user behavior sequence and feature interactions deep interest and machines network(DIMN).We conduct extensive comparison experiments on three public datasets and one private music dataset,which are more recognized in the industry,and the results show that the DIMN obtains a better performance compared with the classical CTR prediction model.展开更多
A groundbreaking method is introduced to leverage machine learn-ing algorithms to revolutionize the prediction of success rates for science fiction films.In the captivating world of the film industry,extensive researc...A groundbreaking method is introduced to leverage machine learn-ing algorithms to revolutionize the prediction of success rates for science fiction films.In the captivating world of the film industry,extensive research and accurate forecasting are vital to anticipating a movie’s triumph prior to its debut.Our study aims to harness the power of available data to estimate a film’s early success rate.With the vast resources offered by the internet,we can access a plethora of movie-related information,including actors,directors,critic reviews,user reviews,ratings,writers,budgets,genres,Facebook likes,YouTube views for movie trailers,and Twitter followers.The first few weeks of a film’s release are crucial in determining its fate,and online reviews and film evaluations profoundly impact its opening-week earnings.Hence,our research employs advanced supervised machine learning techniques to predict a film’s triumph.The Internet Movie Database(IMDb)is a comprehensive data repository for nearly all movies.A robust predictive classification approach is developed by employing various machine learning algorithms,such as fine,medium,coarse,cosine,cubic,and weighted KNN.To determine the best model,the performance of each feature was evaluated based on composite metrics.Moreover,the significant influences of social media platforms were recognized including Twitter,Instagram,and Facebook on shaping individuals’opinions.A hybrid success rating prediction model is obtained by integrating the proposed prediction models with sentiment analysis from available platforms.The findings of this study demonstrate that the chosen algorithms offer more precise estimations,faster execution times,and higher accuracy rates when compared to previous research.By integrating the features of existing prediction models and social media sentiment analysis models,our proposed approach provides a remarkably accurate prediction of a movie’s success.This breakthrough can help movie producers and marketers anticipate a film’s triumph before its release,allowing them to tailor their promotional activities accordingly.Furthermore,the adopted research lays the foundation for developing even more accurate prediction models,considering the ever-increasing significance of social media platforms in shaping individ-uals’opinions.In conclusion,this study showcases the immense potential of machine learning algorithms in predicting the success rate of science fiction films,opening new avenues for the film industry.展开更多
This study examines whether a group of captive false killer whales(P seudorca crassidens) showed variations in the vocal rate around feeding times. The high level of motivation to express appetitive behaviors in capti...This study examines whether a group of captive false killer whales(P seudorca crassidens) showed variations in the vocal rate around feeding times. The high level of motivation to express appetitive behaviors in captive animals may lead them to respond with changes of the behavioral activities during the time prior to food deliveries which are referred to as food anticipatory activity. False killer whales at Qingdao Polar Ocean World(Qingdao, China) showed signifi cant variations of the rates of both the total sounds and sound classes(whistles, clicks, and burst pulses) around feedings. Precisely, from the Transition interval that recorded the lowest vocalization rate(3.40 s/m/d), the whales increased their acoustic emissions upon trainers' arrival(13.08 s/m/d). The high rate was maintained or intensifi ed throughout the food delivery(25.12 s/m/d), and then reduced immediately after the animals were fed(9.91 s/m/d). These changes in the false killer whales sound production rates around feeding times supports the hypothesis of the presence of a food anticipatory vocal activity. Although sound rates may not give detailed information regarding referential aspects of the animal communication it might still shed light about the arousal levels of the individuals during different social or environmental conditions. Further experiments should be performed to assess if variations of the time of feeding routines may affect the vocal activity of cetaceans in captivity as well as their welfare.展开更多
In recent years,a series of major natural gas exploration discoveries and breakthroughs have been achieved in deep and ultra-deep carbonate gas reservoirs in the Sichuan Basin,and all discovered gas reservoirs are cha...In recent years,a series of major natural gas exploration discoveries and breakthroughs have been achieved in deep and ultra-deep carbonate gas reservoirs in the Sichuan Basin,and all discovered gas reservoirs are characterized by great burial depth,complex pore structures and high formation temperature and pore pressure.In order to accurately predict the gas flow rate of single well in high temperature and high pressure(HTHP)gas reservoirs and clarify the gas flow characteristics under formation conditions,this paper establishes a productivity simulation experimental device and method based on the formation temperature and pore pressure of carbonate gas reservoirs in the Middle Permian Qixia Formation of northwestern Sichuan Basin and the Upper Sinian Dengying Formation of central Sichuan Basin.Then,the cores of above mentioned gas reservoirs are selected to carry out the productivity simulation experiment under HTHP.Finally,the gas flow characteristics are studied.And the following research results are obtained.First,the newly established productivity simulation experimental device and method suitable for the conditions of 160℃ formation temperature and 100 MPa pore pressure is used to predict the natural gas AOF(absolute open flow)of Well S-1 in the Qixia Formation gas reservoir of northwestern Sichuan Basin.And the prediction result is better accordant with the calculation result of theoretical model,with a relative error of only 2.12%.Second,based on the Klinkenberg permeability under surface conditions,the single-well gas flow rate calculated from the productivity simulation experiment is better accordant with the gas flow rate from field completion testing;while based on the Klinkenberg permeability under formation conditions,the single-well gas flow rate calculated from the productivity simulation experiment is better accordant with the AOF.Third,the change of formation temperature and pore pressure has a significant influence on rock permeability,and the permeability is more sensitive to stress than to temperature.Fourth,to carry out the reservoir stress sensitivity experiment and the productivity simulation experiment,it is required that core samples be recovered to the formation conditions for aging,or the experimental results may have characteristics of strong stress sensitivity and cannot be used for reservoir engineering evaluation directly.In conclusion,the production rate and AOF of HTHP gas wells can be predicted accurately by means of productivity simulation experiment,based on drilling core samples.In addition,the Klinkenberg permeability under formation conditions can be evaluated by using the relational expression between surface or Klinkenberg permeability under formation conditions and single-well gas flow rate,combined with gas well testing data.展开更多
Purpose-When recommending products to consumers,it is important to be able to accurately predict how consumers will rate them.However,existing collaborative filtering models are difficult to achieve a balance between ...Purpose-When recommending products to consumers,it is important to be able to accurately predict how consumers will rate them.However,existing collaborative filtering models are difficult to achieve a balance between rating prediction accuracy and complexity.Therefore,the purpose of this paper is to propose an accurate and effective model to predict users’ratings of products for the accurate recommendation of products to users.Design/methodology/approach-First,we introduce an attention mechanism that dynamically assigns weights to user preferences,highlighting key interaction information and enhancing the model’s understanding of user behavior.Second,a fold embedding strategy is employed to segment user interaction data,increasing the information density of each subset while reducing the complexity of the attention mechanism.Finally,a masking strategy is integrated to mitigate overfitting by concealing portions of user-item interactions,thereby improving the model’s generalization ability.Findings-The experimental results demonstrate that the proposed model significantly minimizes prediction error across five real-world datasets.On average,the evaluation metrics root mean square error(RMSE)and mean absolute error(MAE)are reduced by 9.11 and 13.3%,respectively.Additionally,the Friedman test results confirm that these improvements are statistically significant.Consequently,the proposed model more accurately captures the intrinsic correlation between users and products,leading to a substantial reduction in prediction error.Originality/value-We propose a novel collaborative filtering model to learn the user-item interaction matrix effectively.Additionally,we introduce a fold embedding strategy to reduce the computational resource consumption of the attention mechanism.Finally,we implement a masking strategy to encourage the model to focus on key features and patterns,thereby mitigating overfitting.展开更多
Autoencoder-based rating prediction methods with external attributes have received wide attention due to their ability to accurately capture users'preferences.However,existing methods still have two significant li...Autoencoder-based rating prediction methods with external attributes have received wide attention due to their ability to accurately capture users'preferences.However,existing methods still have two significant limitations:i)External attributes are often unavailable in the real world due to privacy issues,leading to low quality of representations;and ii)existing methods lack considering complex associations in users'rating behaviors during the encoding process.To meet these challenges,this paper innovatively proposes an inherent-attribute-aware dual-graph autoencoder,named IADGAE,for rating prediction.To address the low quality of representations due to the unavailability of external attributes,we propose an inherent attribute perception module that mines inductive user active patterns and item popularity patterns from users'rating behaviors to strengthen user and item representations.To exploit the complex associations hidden in users’rating behaviors,we design an encoder on the item-item co-occurrence graph to capture the co-occurrence frequency features among items.Moreover,we propose a dual-graph feature encoder framework to simultaneously encode and fuse the high-order representations learned from the user-item rating graph and item-item co-occurrence graph.Extensive experiments on three real datasets demonstrate that IADGAE is effective and outperforms existing rating prediction methods,which achieves a significant improvement of 4.51%~41.63%in the RMSE metric.展开更多
Prediction of hepatitis B surface antigen(HBsAg)decline rates during treatment is crucial for achieving a higher proportion of functional cure outcomes in patients with chronic hepatitis B(CHB),and so is the identific...Prediction of hepatitis B surface antigen(HBsAg)decline rates during treatment is crucial for achieving a higher proportion of functional cure outcomes in patients with chronic hepatitis B(CHB),and so is the identification of favorable patients.A total of 371 patients who received pegylated interferon alpha monotherapy or sequential/combined nucleos(t)ide analogues therapy between May 2018 and July 2024 were included for follow-up analysis.The patients were divided into a training set,a validation set and a test set via time series partitioning and random partitioning methods.The primary outcome was the prediction of HBsAg decline rate at each medical visit via linear mixed effects model.Patient stratification was secondary outcomes assessed using group-based trajectory model.The cumulative number of functional cures among 371 patients was 76(20%,95%CI:16%-25%).Three groups,namely rapid high-clearance,delayed high-clearance,and slow low-clearance,were identified by the group trajectory model.The overall accuracy of the time-plus-group dual-effect prediction model was 84%(95%CI:81%-87%),which was approximately 10%higher than that of the time-effect prediction model after 24 weeks of treatment.When the computational cost was combined,a pragmatic prediction strategy with robust individual prediction performance was obtained.The constructed group trajectory model and prediction strategy may have the potential to dynamically identify favorable patients and dynamically predict the HBsAg decline rate,thereby improving the functional cure rate in clinical practice.展开更多
To validate the feasibility of using near-infrared(NIR)spectroscopy for real-time monitoring of multiple active pharmaceutical ingredients dissolution,this study focused on Guizhi Fuling capsules and tablets.The NIR s...To validate the feasibility of using near-infrared(NIR)spectroscopy for real-time monitoring of multiple active pharmaceutical ingredients dissolution,this study focused on Guizhi Fuling capsules and tablets.The NIR spectroscopy fiber probe was inserted into the dissolution apparatus and connected to a Fourier transform near-infrared spectrometer(FT-NIR)to capture spectral data.During the dissolution tests,dissolution behavior curves for seven components,gallic acid(GA),alibiflorin(ALI),paeoniflorin(PF),paeonol(PAE),amygdalin(AMY),cinnamaldehyde(CL),and cinnamic acid(CA)in the capsules,were obtained by sampling from the dissolution cups at specific time intervals.Linear regression was applied to models corrected using various pre-process techniques with the partial least squares(PLS)algorithm.Additionally,an artificial neural network(ANN),a nonlinear regression algorithm,was utilized to explore the complex relationship between spectra and multicomponent dissolution.Ultimately,the ANN model achieved a lower prediction mean square error(RMSEP)and relative error compared to the PLS model,with significantly higher correlation coefficient(Rp)for the validation set.The highest Rpvalue reached 0.8825.The paired t-test results also indicated no significant difference between predicted and measured values.Furthermore,the ANN model demonstrated the best predictive performance in the tablet experiments,achieving an Rpof 0.8134.The findings indicate that real-time monitoring of multicomponent drug dissolution using NIR spectroscopy combined with chemometric methods is feasible,offering a promising new direction to replace traditional dissolution testing.展开更多
基金supported in part by National Science Foundation of China under Grants No.61303105 and 61402304the Humanity&Social Science general project of Ministry of Education under Grants No.14YJAZH046+2 种基金the Beijing Natural Science Foundation under Grants No.4154065the Beijing Educational Committee Science and Technology Development Planned under Grants No.KM201410028017Academic Degree Graduate Courses group projects
文摘Collaborative Filtering(CF) is a leading approach to build recommender systems which has gained considerable development and popularity. A predominant approach to CF is rating prediction recommender algorithm, aiming to predict a user's rating for those items which were not rated yet by the user. However, with the increasing number of items and users, thedata is sparse.It is difficult to detectlatent closely relation among the items or users for predicting the user behaviors. In this paper,we enhance the rating prediction approach leading to substantial improvement of prediction accuracy by categorizing according to the genres of movies. Then the probabilities that users are interested in the genres are computed to integrate the prediction of each genre cluster. A novel probabilistic approach based on the sentiment analysis of the user reviews is also proposed to give intuitional explanations of why an item is recommended.To test the novel recommendation approach, a new corpus of user reviews on movies obtained from the Internet Movies Database(IMDB) has been generated. Experimental results show that the proposed framework is effective and achieves a better prediction performance.
基金Supported by the Scientific Research Foundation of Liaoning Provincial Department of Education(No.LJKZ0139).
文摘This paper examines the prediction of film ratings.Firstly,in the data feature engineering,feature construction is performed based on the original features of the film dataset.Secondly,the clustering algorithm is utilized to remove singular film samples,and feature selections are carried out.When solving the problem that film samples of the target domain are unlabelled,it is impossible to train a model and address the inconsistency in the feature dimension for film samples from the source domain.Therefore,the domain adaptive transfer learning model combined with dimensionality reduction algorithms is adopted in this paper.At the same time,in order to reduce the prediction error of models,the stacking ensemble learning model for regression is also used.Finally,through comparative experiments,the effectiveness of the proposed method is verified,which proves to be better predicting film ratings in the target domain.
基金financially supported by CNOOC China Co., Ltd. Zhanjiang Branch (CNOOC-KJ135ZDXM3 8ZJ05ZJ)。
文摘Fossil fuels are undoubtedly important, and drilling technology plays an important role in realizing fossil fuel exploration;therefore, the prediction and evaluation of drilling efficiency is a key research goal in the industry. Limited by the unknown geological environment and complex operating procedures, the prediction and evaluation of drilling efficiency were very difficult before the introduction of machine learning algorithms. This review statistically analyses rate of penetration(ROP) prediction models established based on machine learning algorithms;establishes an overall framework including data collection, data preprocessing, model establishment, and accuracy evaluation;and compares the effectiveness of different algorithms in each link of the process. This review also compares the prediction accuracy of different machine learning models and traditional models commonly used in this field and demonstrates that machine learning models are the most effective technical means in current ROP prediction modeling.
基金This work is supported by National Natural Science Foundation of China(No.61672133)Sichuan Science and Technology Program(No.2019YFG0535)the 111 Project(No.B17008).
文摘Recommender systems are rapidly transforming the digital world into intelligent information hubs.The valuable context information associated with the users’prior transactions has played a vital role in determining the user preferences for items or rating prediction.It has been a hot research topic in collaborative filtering-based recommender systems for the last two decades.This paper presents a novel Context Based Rating Prediction(CBRP)model with a unique similarity scoring estimation method.The proposed algorithm computes a context score for each candidate user to construct a similarity pool for the given subject user-item pair and intuitively choose the highly influential users to forecast the item ratings.The context scoring strategy has an inherent capability to incorporate multiple conditional factors to filter down the most relevant recommendations.Compared with traditional similarity estimation methods,CBRP makes it possible for the full use of neighboring collaborators’choice on various conditions.We conduct experiments on three publicly available datasets to evaluate our proposed method with random user-item pairs and got considerable improvement in prediction accuracy over the standard evaluation measures.Also,we evaluate prediction accuracy for every user-item pair in the system and the results show that our proposed framework has outperformed existing methods.
基金“Peso-Dollar Exchange Rate Prediction:Fundamentals vs.Internet Search Indicators”,which was sponsored by the Universidad Autonoma de Nuevo Leon,Mexico(PAICYT Project#375-CSA-2022).
文摘Recently,internet users have significantly increased their use of search engines,and market investors are no exception.As a result,predictive models that incorporate scattered web-based information are developing as an area of forecasting.The objective of this research is to compare the predictive accuracy of fundamental macroeconomic variables,online attention series measured by the Google Trends search volume index,and a combination of both data types for the Mexican,Brazilian,Chilean,and Colombian currencies paired with the USD.The exchange rate series used in this study are sourced from a real-time platform.Four indicators capturing the fundamental macroeconomic differences between these emerging economies and the U.S.from January 2004 to March 2021(monthly)were analyzed.To assess the predictive performance of the KNN algorithm,OLS regression and the random walk with drift model were compared.Considering in-sample predictions,the results generally exhibit lower estimation errors in the random walk with drift model,but in the joint fundamental–online attention data,the KNN and OLS predictions are more accurate than those of the random walk with drift.However,the KNN predictions based on out-of-sample fit generate the lowest estimation errors and the most accurate predictions for the joint fundamental–online attention data.Additionally,performance testing indicates that the KNN extended model outperforms the out-ofsample forecast for the OLS regression and the random walk with drift model.
基金funded by the National Natural Science Foundation of China(No.72072091)the Natural Science Foundation of Colleges and Universities of Jiangsu Province(Nos.21KJA520002 and 22KJA520005)+1 种基金the Ministry of Education Supply and Demand Matching Employment and Education Project(No.2023121801795)the Postgraduate Research&Practice Innovation Program of Jiangsu Province(Nos.KYCX23_2345 and SJCX24_1164).
文摘The effectiveness of recommendation systems heavily relies on accurately predicting user ratings for items based on user preferences and item attributes derived from ratings and reviews.However,the increasing presence of fake user data in these ratings and reviews poses significant challenges,hindering feature extraction,diminishing rating prediction accuracy,and eroding user trust in the system.To tackle this issue,we propose a robust rating prediction model for recommendation systems that integrates fake user detection and multi-layer feature fusion.Our model utilizes a GraphSAGE-based submodel to filter out fake user data from rating data and review texts.To strengthen fake user detection,we enhance GraphSAGE by selecting aggregation neighbors based on the collusion fraud degree among users,and employ an attention mechanism to weigh the contribution of each neighbor during representation aggregation.Furthermore,we introduce a multi-layer feature fusion submodel to integrate diverse features extracted from the filtered ratings and reviews.For deep feature extraction from review texts,we implement a temporal attention mechanism to analyze the relevance of reviews over time.For shallow feature extraction from rating data,we incorporate trust evaluation mechanism and cloud model to assess the influence of trusted neighbors’ratings.In our evaluation,we compare our model against six baseline models for fake user detection and four rating prediction models across five datasets.The results demonstrate that our model exhibits significant performance advantages in both fake user detection and rating prediction.
基金National Natural Science Foundation of China (51175502)
文摘Sensor selection and optimization is one of the important parts in design for testability. To address the problems that the traditional sensor optimization selection model does not take the requirements of prognostics and health management especially fault prognostics for testability into account and does not consider the impacts of sensor actual attributes on fault detectability, a novel sensor optimization selection model is proposed. Firstly, a universal architecture for sensor selection and optimization is provided. Secondly, a new testability index named fault predictable rate is defined to describe fault prognostics requirements for testability. Thirdly, a sensor selection and optimization model for prognostics and health management is constructed, which takes sensor cost as objective function and the defined testability indexes as constraint conditions. Due to NP-hard property of the model, a generic algorithm is designed to obtain the optimal solution. At last, a case study is presented to demonstrate the sensor selection approach for a stable tracking servo platform. The application results and comparison analysis show the proposed model and algorithm are effective and feasible. This approach can be used to select sensors for prognostics and health management of any system.
基金Project(2010CB732004)supported by the National Basic Research Program of ChinaProjects(50934006,41272304)supported by the National Natural Science Foundation of China
文摘Rate of penetration(ROP) of a tunnel boring machine(TBM) in a rock environment is generally a key parameter for the successful accomplishment of a tunneling project. The objectives of this work are to compare the accuracy of prediction models employing partial least squares(PLS) regression and support vector machine(SVM) regression technique for modeling the penetration rate of TBM. To develop the proposed models, the database that is composed of intact rock properties including uniaxial compressive strength(UCS), Brazilian tensile strength(BTS), and peak slope index(PSI), and also rock mass properties including distance between planes of weakness(DPW) and the alpha angle(α) are input as dependent variables and the measured ROP is chosen as an independent variable. Two hundred sets of data are collected from Queens Water Tunnel and Karaj-Tehran water transfer tunnel TBM project. The accuracy of the prediction models is measured by the coefficient of determination(R2) and root mean squares error(RMSE) between predicted and observed yield employing 10-fold cross-validation schemes. The R2 and RMSE of prediction are 0.8183 and 0.1807 for SVMR method, and 0.9999 and 0.0011 for PLS method, respectively. Comparison between the values of statistical parameters reveals the superiority of the PLSR model over SVMR one.
文摘In this paper, the method which can combine different seismic data with the different precision and completeness, even the palaeo-earthquake data, has been applied to estimate the yearly seismic moment rate in the seismic region. Based on this, the predictable model of regional time-magnitude has been used in North China and Southwest China. The normal correlation between the time interval of the events and the magnitude of the last strong earthquake shows that the model is suitable. The value of the parameter c is less than the average value of 0.33 that is obtained from the events occurred in the plate boundary in the world. It is explained that the correlativity between the recurrence interval of the earthquake and the magnitude of the last strong event is not obvious. It is shown that the continental earthquakes in China are different from that occurred in the plate boundary and the recurrence model for the continental events are different from the one for the plate boundary events. Finally the seismic risk analysis based on this model for North China and Southwest China is given in this paper.
文摘The reliability of electrical connectors has critical impact on electronic systems. It is usually characterized by failure rate prediction value according to standard MIL-HDBK-217(or GJB-299 C in Chinese) in engineering practice. Given to their limitations and mislead results, a new failure rate prediction models needs to be presented. The presented model aims at the mechanism of increase of film thickness which leads to the increase of contact resistance. The estimated failure rate value can be given at different environmental conditions,and some of the factors affecting the reliability are taken into account. Accelerated degradation test(ADT) was conducted on GJB599 III series electrical connector. The failure rate prediction model can be simply formed and convenient to calculate the expression of failure rate changing with time at various temperature and vibration conditions. This model gives an objective assessment in short time, which makes it convenient to be applied to the engineering.
基金Supported by the National Natural Science Foundation of China (62272214)。
文摘In the era of intelligent economy, the click-through rate(CTR) prediction system can evaluate massive service information based on user historical information, and screen out the products that are most likely to be favored by users, thus realizing customized push of information and achieve the ultimate goal of improving economic benefits. Sequence modeling is one of the main research directions of CTR prediction models based on deep learning. The user's general interest hidden in the entire click history and the short-term interest hidden in the recent click behaviors have different influences on the CTR prediction results, which are highly important. In terms of capturing the user's general interest, existing models paid more attention to the relationships between item embedding vectors(point-level), while ignoring the relationships between elements in item embedding vectors(union-level). The Lambda layer-based Convolutional Sequence Embedding(LCSE) model proposed in this paper uses the Lambda layer to capture features from click history through weight distribution, and uses horizontal and vertical filters on this basis to learn the user's general preferences from union-level and point-level. In addition, we also incorporate the user's short-term preferences captured by the embedding-based convolutional model to further improve the prediction results. The AUC(Area Under Curve) values of the LCSE model on the datasets Electronic, Movie & TV and MovieLens are 0.870 7, 0.903 6 and 0.946 7, improving 0.45%, 0.36% and 0.07% over the Caser model, proving the effectiveness of our proposed model.
文摘This research concentrates to model an efficient thyroid prediction approach,which is considered a baseline for significant problems faced by the women community.The major research problem is the lack of automated model to attain earlier prediction.Some existing model fails to give better prediction accuracy.Here,a novel clinical decision support system is framed to make the proper decision during a time of complexity.Multiple stages are followed in the proposed framework,which plays a substantial role in thyroid prediction.These steps include i)data acquisition,ii)outlier prediction,and iii)multi-stage weight-based ensemble learning process(MS-WEL).The weighted analysis of the base classifier and other classifier models helps bridge the gap encountered in one single classifier model.Various classifiers aremerged to handle the issues identified in others and intend to enhance the prediction rate.The proposed model provides superior outcomes and gives good quality prediction rate.The simulation is done in the MATLAB 2020a environment and establishes a better trade-off than various existing approaches.The model gives a prediction accuracy of 97.28%accuracy compared to other models and shows a better trade than others.
文摘In recent years,deep learning has been widely applied in the fields of recommendation systems and click-through rate(CTR)prediction,and thus recommendation models incorporating deep learning have emerged.In addition,the design and implementation of recommendation models using information related to user behavior sequences is an important direction of current research in recommendation systems,and models calculate the likelihood of users clicking on target items based on their behavior sequence information.In order to explore the relationship between features,this paper improves and optimizes on the basis of deep interest network(DIN)proposed by Ali’s team.Based on the user behavioral sequences information,the attentional factorization machine(AFM)is integrated to obtain richer and more accurate behavioral sequence information.In addition,this paper designs a new way of calculating attention weights,which uses the relationship between the cosine similarity of any two vectors and the absolute value of their modal length difference to measure their relevance degree.Thus,a novel deep learning CTR prediction mode is proposed,that is,the CTR prediction network based on user behavior sequence and feature interactions deep interest and machines network(DIMN).We conduct extensive comparison experiments on three public datasets and one private music dataset,which are more recognized in the industry,and the results show that the DIMN obtains a better performance compared with the classical CTR prediction model.
文摘A groundbreaking method is introduced to leverage machine learn-ing algorithms to revolutionize the prediction of success rates for science fiction films.In the captivating world of the film industry,extensive research and accurate forecasting are vital to anticipating a movie’s triumph prior to its debut.Our study aims to harness the power of available data to estimate a film’s early success rate.With the vast resources offered by the internet,we can access a plethora of movie-related information,including actors,directors,critic reviews,user reviews,ratings,writers,budgets,genres,Facebook likes,YouTube views for movie trailers,and Twitter followers.The first few weeks of a film’s release are crucial in determining its fate,and online reviews and film evaluations profoundly impact its opening-week earnings.Hence,our research employs advanced supervised machine learning techniques to predict a film’s triumph.The Internet Movie Database(IMDb)is a comprehensive data repository for nearly all movies.A robust predictive classification approach is developed by employing various machine learning algorithms,such as fine,medium,coarse,cosine,cubic,and weighted KNN.To determine the best model,the performance of each feature was evaluated based on composite metrics.Moreover,the significant influences of social media platforms were recognized including Twitter,Instagram,and Facebook on shaping individuals’opinions.A hybrid success rating prediction model is obtained by integrating the proposed prediction models with sentiment analysis from available platforms.The findings of this study demonstrate that the chosen algorithms offer more precise estimations,faster execution times,and higher accuracy rates when compared to previous research.By integrating the features of existing prediction models and social media sentiment analysis models,our proposed approach provides a remarkably accurate prediction of a movie’s success.This breakthrough can help movie producers and marketers anticipate a film’s triumph before its release,allowing them to tailor their promotional activities accordingly.Furthermore,the adopted research lays the foundation for developing even more accurate prediction models,considering the ever-increasing significance of social media platforms in shaping individ-uals’opinions.In conclusion,this study showcases the immense potential of machine learning algorithms in predicting the success rate of science fiction films,opening new avenues for the film industry.
基金Supported by grants from the Institute of Hydrobiology,Chinese Academy of Sciences
文摘This study examines whether a group of captive false killer whales(P seudorca crassidens) showed variations in the vocal rate around feeding times. The high level of motivation to express appetitive behaviors in captive animals may lead them to respond with changes of the behavioral activities during the time prior to food deliveries which are referred to as food anticipatory activity. False killer whales at Qingdao Polar Ocean World(Qingdao, China) showed signifi cant variations of the rates of both the total sounds and sound classes(whistles, clicks, and burst pulses) around feedings. Precisely, from the Transition interval that recorded the lowest vocalization rate(3.40 s/m/d), the whales increased their acoustic emissions upon trainers' arrival(13.08 s/m/d). The high rate was maintained or intensifi ed throughout the food delivery(25.12 s/m/d), and then reduced immediately after the animals were fed(9.91 s/m/d). These changes in the false killer whales sound production rates around feeding times supports the hypothesis of the presence of a food anticipatory vocal activity. Although sound rates may not give detailed information regarding referential aspects of the animal communication it might still shed light about the arousal levels of the individuals during different social or environmental conditions. Further experiments should be performed to assess if variations of the time of feeding routines may affect the vocal activity of cetaceans in captivity as well as their welfare.
基金Project supported by the Scientific Research and Technology Development Project of PetroChina Company Limited“New Experimental Technology Development of Key Laboratory of Carbonate Reservoir”(No.:2018D-5006-35).
文摘In recent years,a series of major natural gas exploration discoveries and breakthroughs have been achieved in deep and ultra-deep carbonate gas reservoirs in the Sichuan Basin,and all discovered gas reservoirs are characterized by great burial depth,complex pore structures and high formation temperature and pore pressure.In order to accurately predict the gas flow rate of single well in high temperature and high pressure(HTHP)gas reservoirs and clarify the gas flow characteristics under formation conditions,this paper establishes a productivity simulation experimental device and method based on the formation temperature and pore pressure of carbonate gas reservoirs in the Middle Permian Qixia Formation of northwestern Sichuan Basin and the Upper Sinian Dengying Formation of central Sichuan Basin.Then,the cores of above mentioned gas reservoirs are selected to carry out the productivity simulation experiment under HTHP.Finally,the gas flow characteristics are studied.And the following research results are obtained.First,the newly established productivity simulation experimental device and method suitable for the conditions of 160℃ formation temperature and 100 MPa pore pressure is used to predict the natural gas AOF(absolute open flow)of Well S-1 in the Qixia Formation gas reservoir of northwestern Sichuan Basin.And the prediction result is better accordant with the calculation result of theoretical model,with a relative error of only 2.12%.Second,based on the Klinkenberg permeability under surface conditions,the single-well gas flow rate calculated from the productivity simulation experiment is better accordant with the gas flow rate from field completion testing;while based on the Klinkenberg permeability under formation conditions,the single-well gas flow rate calculated from the productivity simulation experiment is better accordant with the AOF.Third,the change of formation temperature and pore pressure has a significant influence on rock permeability,and the permeability is more sensitive to stress than to temperature.Fourth,to carry out the reservoir stress sensitivity experiment and the productivity simulation experiment,it is required that core samples be recovered to the formation conditions for aging,or the experimental results may have characteristics of strong stress sensitivity and cannot be used for reservoir engineering evaluation directly.In conclusion,the production rate and AOF of HTHP gas wells can be predicted accurately by means of productivity simulation experiment,based on drilling core samples.In addition,the Klinkenberg permeability under formation conditions can be evaluated by using the relational expression between surface or Klinkenberg permeability under formation conditions and single-well gas flow rate,combined with gas well testing data.
文摘Purpose-When recommending products to consumers,it is important to be able to accurately predict how consumers will rate them.However,existing collaborative filtering models are difficult to achieve a balance between rating prediction accuracy and complexity.Therefore,the purpose of this paper is to propose an accurate and effective model to predict users’ratings of products for the accurate recommendation of products to users.Design/methodology/approach-First,we introduce an attention mechanism that dynamically assigns weights to user preferences,highlighting key interaction information and enhancing the model’s understanding of user behavior.Second,a fold embedding strategy is employed to segment user interaction data,increasing the information density of each subset while reducing the complexity of the attention mechanism.Finally,a masking strategy is integrated to mitigate overfitting by concealing portions of user-item interactions,thereby improving the model’s generalization ability.Findings-The experimental results demonstrate that the proposed model significantly minimizes prediction error across five real-world datasets.On average,the evaluation metrics root mean square error(RMSE)and mean absolute error(MAE)are reduced by 9.11 and 13.3%,respectively.Additionally,the Friedman test results confirm that these improvements are statistically significant.Consequently,the proposed model more accurately captures the intrinsic correlation between users and products,leading to a substantial reduction in prediction error.Originality/value-We propose a novel collaborative filtering model to learn the user-item interaction matrix effectively.Additionally,we introduce a fold embedding strategy to reduce the computational resource consumption of the attention mechanism.Finally,we implement a masking strategy to encourage the model to focus on key features and patterns,thereby mitigating overfitting.
基金supported in part by National Natural Science Foundation of China(U21B2015,61972300)in part by Young Scientists Fund of the National Natural Science Foundation of China(62202356)+1 种基金in part by Young Talent Fund of Association for Science and Technology in Shaanxi(20220113)in part by Intelligent Financial Software Engineering New Technology Joint Laboratory Project(99901220858)。
文摘Autoencoder-based rating prediction methods with external attributes have received wide attention due to their ability to accurately capture users'preferences.However,existing methods still have two significant limitations:i)External attributes are often unavailable in the real world due to privacy issues,leading to low quality of representations;and ii)existing methods lack considering complex associations in users'rating behaviors during the encoding process.To meet these challenges,this paper innovatively proposes an inherent-attribute-aware dual-graph autoencoder,named IADGAE,for rating prediction.To address the low quality of representations due to the unavailability of external attributes,we propose an inherent attribute perception module that mines inductive user active patterns and item popularity patterns from users'rating behaviors to strengthen user and item representations.To exploit the complex associations hidden in users’rating behaviors,we design an encoder on the item-item co-occurrence graph to capture the co-occurrence frequency features among items.Moreover,we propose a dual-graph feature encoder framework to simultaneously encode and fuse the high-order representations learned from the user-item rating graph and item-item co-occurrence graph.Extensive experiments on three real datasets demonstrate that IADGAE is effective and outperforms existing rating prediction methods,which achieves a significant improvement of 4.51%~41.63%in the RMSE metric.
基金Supported by the National Natural Science Foundation of China[12171396,12471473,and 32171152].
文摘Prediction of hepatitis B surface antigen(HBsAg)decline rates during treatment is crucial for achieving a higher proportion of functional cure outcomes in patients with chronic hepatitis B(CHB),and so is the identification of favorable patients.A total of 371 patients who received pegylated interferon alpha monotherapy or sequential/combined nucleos(t)ide analogues therapy between May 2018 and July 2024 were included for follow-up analysis.The patients were divided into a training set,a validation set and a test set via time series partitioning and random partitioning methods.The primary outcome was the prediction of HBsAg decline rate at each medical visit via linear mixed effects model.Patient stratification was secondary outcomes assessed using group-based trajectory model.The cumulative number of functional cures among 371 patients was 76(20%,95%CI:16%-25%).Three groups,namely rapid high-clearance,delayed high-clearance,and slow low-clearance,were identified by the group trajectory model.The overall accuracy of the time-plus-group dual-effect prediction model was 84%(95%CI:81%-87%),which was approximately 10%higher than that of the time-effect prediction model after 24 weeks of treatment.When the computational cost was combined,a pragmatic prediction strategy with robust individual prediction performance was obtained.The constructed group trajectory model and prediction strategy may have the potential to dynamically identify favorable patients and dynamically predict the HBsAg decline rate,thereby improving the functional cure rate in clinical practice.
基金support of the Open Fund from the State Key Laboratory of New-tech for Chinese Medicine Pharmaceutical Process(SKL2020Z0203)Science and Technology Project of Haihe Laboratory of Modern Chinese Medicine(22HHZYSS00004)the 14th University Student Science and Technology Innovation Fund of Tianjin University of Traditional Chinese Medicine(KJ12)。
文摘To validate the feasibility of using near-infrared(NIR)spectroscopy for real-time monitoring of multiple active pharmaceutical ingredients dissolution,this study focused on Guizhi Fuling capsules and tablets.The NIR spectroscopy fiber probe was inserted into the dissolution apparatus and connected to a Fourier transform near-infrared spectrometer(FT-NIR)to capture spectral data.During the dissolution tests,dissolution behavior curves for seven components,gallic acid(GA),alibiflorin(ALI),paeoniflorin(PF),paeonol(PAE),amygdalin(AMY),cinnamaldehyde(CL),and cinnamic acid(CA)in the capsules,were obtained by sampling from the dissolution cups at specific time intervals.Linear regression was applied to models corrected using various pre-process techniques with the partial least squares(PLS)algorithm.Additionally,an artificial neural network(ANN),a nonlinear regression algorithm,was utilized to explore the complex relationship between spectra and multicomponent dissolution.Ultimately,the ANN model achieved a lower prediction mean square error(RMSEP)and relative error compared to the PLS model,with significantly higher correlation coefficient(Rp)for the validation set.The highest Rpvalue reached 0.8825.The paired t-test results also indicated no significant difference between predicted and measured values.Furthermore,the ANN model demonstrated the best predictive performance in the tablet experiments,achieving an Rpof 0.8134.The findings indicate that real-time monitoring of multicomponent drug dissolution using NIR spectroscopy combined with chemometric methods is feasible,offering a promising new direction to replace traditional dissolution testing.
