Root zone soil moisture at one and two meter depths are forecasted four days into the future. In this article, we propose a new multivariate output prediction approach to root zone soil moisture assessment using learn...Root zone soil moisture at one and two meter depths are forecasted four days into the future. In this article, we propose a new multivariate output prediction approach to root zone soil moisture assessment using learning machine models. These models are known for their robustness, efficiency, and sparseness;they provide a statistically sound approach to solving the inverse problem and thus to building statistical models. The multivariate relevance vector machine (MVRVM) is used to build a model that forecasts soil moisture states based upon current soil moisture and soil temperature conditions. The methodology combines the data at different depths from 5 cm to 50 cm, the largest of which corresponds to the depth at which the soil moisture sensors are generally operational, to produce soil moisture predictions at larger depths. The MVRVM test results for soil moisture predictions at 1 m and 2 m depth on the 4th day are excellent with RMSE = 0.0131 m3/m3 for 1 m;and RMSE = 0.0015 m3/m3 for 2 m forecasted values. The statistics of predictions for 4th day (CoE = 0.87 for 1 m and CoE = 0.96 for 2 m) indicate good model generalization capability and computations show good agreement with actual measurements with R2 = 0.88 and R2 = 0.97 for 1 m and 2 m depths, respectively. The MVRVM produces good results for all four days. Bootstrapping is used to check over/under-fitting and uncertainty in model estimates.展开更多
The software engineering technique makes it possible to create high-quality software.One of the most significant qualities of good software is that it is devoid of bugs.One of the most time-consuming and costly softwar...The software engineering technique makes it possible to create high-quality software.One of the most significant qualities of good software is that it is devoid of bugs.One of the most time-consuming and costly software proce-dures isfinding andfixing bugs.Although it is impossible to eradicate all bugs,it is feasible to reduce the number of bugs and their negative effects.To broaden the scope of bug prediction techniques and increase software quality,numerous causes of software problems must be identified,and successful bug prediction models must be implemented.This study employs a hybrid of Faster Convolution Neural Network and the Moth Flame Optimization(MFO)algorithm to forecast the number of bugs in software based on the program data itself,such as the line quantity in codes,methods characteristics,and other essential software aspects.Here,the MFO method is used to train the neural network to identify optimal weights.The proposed MFO-FCNN technique is compared with existing methods such as AdaBoost(AB),Random Forest(RF),K-Nearest Neighbour(KNN),K-Means Clustering(KMC),Support Vector Machine(SVM)and Bagging Clas-sifier(BC)are examples of machine learning(ML)techniques.The assessment method revealed that machine learning techniques may be employed successfully and through a high level of accuracy.The obtained data revealed that the proposed strategy outperforms the traditional approach.展开更多
Bead sttape in underwater rotating arc welding was affected by several welding parameters. RVM （ relevance vector machine） was used to build a model to predict weld bead shape. The training data set of RVM eortsists...Bead sttape in underwater rotating arc welding was affected by several welding parameters. RVM （ relevance vector machine） was used to build a model to predict weld bead shape. The training data set of RVM eortsists of the welding parameters which are rotational frequency, rotational radius, height of torch and welding current and the features of the bead shape. The maximum error and mean error for prediction of width are 0. 10 mm and 0. 09 mm, respectively, and the maximum error and mean error for prediction of penetration are 0. 31 mm and 0. 12mm, respectively, which are showed that the prediction model can achieve higher prediction precision at reasonably small size of training data set.展开更多
文摘Root zone soil moisture at one and two meter depths are forecasted four days into the future. In this article, we propose a new multivariate output prediction approach to root zone soil moisture assessment using learning machine models. These models are known for their robustness, efficiency, and sparseness;they provide a statistically sound approach to solving the inverse problem and thus to building statistical models. The multivariate relevance vector machine (MVRVM) is used to build a model that forecasts soil moisture states based upon current soil moisture and soil temperature conditions. The methodology combines the data at different depths from 5 cm to 50 cm, the largest of which corresponds to the depth at which the soil moisture sensors are generally operational, to produce soil moisture predictions at larger depths. The MVRVM test results for soil moisture predictions at 1 m and 2 m depth on the 4th day are excellent with RMSE = 0.0131 m3/m3 for 1 m;and RMSE = 0.0015 m3/m3 for 2 m forecasted values. The statistics of predictions for 4th day (CoE = 0.87 for 1 m and CoE = 0.96 for 2 m) indicate good model generalization capability and computations show good agreement with actual measurements with R2 = 0.88 and R2 = 0.97 for 1 m and 2 m depths, respectively. The MVRVM produces good results for all four days. Bootstrapping is used to check over/under-fitting and uncertainty in model estimates.
文摘The software engineering technique makes it possible to create high-quality software.One of the most significant qualities of good software is that it is devoid of bugs.One of the most time-consuming and costly software proce-dures isfinding andfixing bugs.Although it is impossible to eradicate all bugs,it is feasible to reduce the number of bugs and their negative effects.To broaden the scope of bug prediction techniques and increase software quality,numerous causes of software problems must be identified,and successful bug prediction models must be implemented.This study employs a hybrid of Faster Convolution Neural Network and the Moth Flame Optimization(MFO)algorithm to forecast the number of bugs in software based on the program data itself,such as the line quantity in codes,methods characteristics,and other essential software aspects.Here,the MFO method is used to train the neural network to identify optimal weights.The proposed MFO-FCNN technique is compared with existing methods such as AdaBoost(AB),Random Forest(RF),K-Nearest Neighbour(KNN),K-Means Clustering(KMC),Support Vector Machine(SVM)and Bagging Clas-sifier(BC)are examples of machine learning(ML)techniques.The assessment method revealed that machine learning techniques may be employed successfully and through a high level of accuracy.The obtained data revealed that the proposed strategy outperforms the traditional approach.
基金The authors wish to thank the financial support for this research from National Natural Science Foundation of China （ No. 50705030） , Natural Science Foundaiion of Guangdong Province of China （No. 9151008019000008 ） and the Fundamental Research Funds for the Central Universities （No. 2009ZM0318）.
文摘Bead sttape in underwater rotating arc welding was affected by several welding parameters. RVM （ relevance vector machine） was used to build a model to predict weld bead shape. The training data set of RVM eortsists of the welding parameters which are rotational frequency, rotational radius, height of torch and welding current and the features of the bead shape. The maximum error and mean error for prediction of width are 0. 10 mm and 0. 09 mm, respectively, and the maximum error and mean error for prediction of penetration are 0. 31 mm and 0. 12mm, respectively, which are showed that the prediction model can achieve higher prediction precision at reasonably small size of training data set.
