Click-through information is considered as a valuable source of users＇ implicit relevance feedback. As user behavior is usually influenced by a number of factors such as position, presentation style and site reputati...Click-through information is considered as a valuable source of users＇ implicit relevance feedback. As user behavior is usually influenced by a number of factors such as position, presentation style and site reputation, researchers have proposed a variety of assumptions to generate a reasonable estimation of result relevance. Therefore, many click models have been proposed to describe how user click action happens and to predict click probability （and search result relevance）. This work builds upon many years of existing efforts from THUIR labs, summarizes the most recent advances and provides a series of practical click models. In this paper, we give an introduction of how to build an effective click model. We use two click models as specific examples to introduce the general procedures of building a click model. We also introduce common evaluation metrics for the comparison of different click models. Some useful datasets and tools are also introduced to help readers better understand and implement existing click models. The goal of this survey is to bring together current efforts in the area, summarize the research performed so far and give a view on building click models for web search.展开更多
Modem search engines record user interactions and use them to improve search quality. In particular, user click-through has been successfully used to improve click- through rate （CTR）, Web search ranking, and query ...Modem search engines record user interactions and use them to improve search quality. In particular, user click-through has been successfully used to improve click- through rate （CTR）, Web search ranking, and query rec- ommendations and suggestions. Although click-through logs can provide implicit feedback of users＇ click preferences, de- riving accurate absolute relevance judgments is difficult be- cause of the existence of click noises and behavior biases. Previous studies showed that user clicking behaviors are bi- ased toward many aspects such as ＂position＂ （user＇s attention decreases from top to bottom） and ＂trust＂ （Web site reputa- tions will affect user＇s judgment）. To address these problems, researchers have proposed several behavior models （usually referred to as click models） to describe users？ practical browsing behaviors and to obtain an unbiased estimation of result relevance. In this study, we review recent efforts to construct click models for better search ranking and propose a novel convolutional neural network architecture for build- ing click models. Compared to traditional click models, our model not only considers user behavior assumptions as input signals but also uses the content and context information of search engine result pages. In addition, our model uses pa- rameters from traditional click models to restrict the meaning of some outputs in our model＇s hidden layer. Experimental results show that the proposed model can achieve consider- able improvement over state-of-the-art click models based on the evaluation metric of click perplexity.展开更多
文摘Click-through information is considered as a valuable source of users＇ implicit relevance feedback. As user behavior is usually influenced by a number of factors such as position, presentation style and site reputation, researchers have proposed a variety of assumptions to generate a reasonable estimation of result relevance. Therefore, many click models have been proposed to describe how user click action happens and to predict click probability （and search result relevance）. This work builds upon many years of existing efforts from THUIR labs, summarizes the most recent advances and provides a series of practical click models. In this paper, we give an introduction of how to build an effective click model. We use two click models as specific examples to introduce the general procedures of building a click model. We also introduce common evaluation metrics for the comparison of different click models. Some useful datasets and tools are also introduced to help readers better understand and implement existing click models. The goal of this survey is to bring together current efforts in the area, summarize the research performed so far and give a view on building click models for web search.
文摘Modem search engines record user interactions and use them to improve search quality. In particular, user click-through has been successfully used to improve click- through rate （CTR）, Web search ranking, and query rec- ommendations and suggestions. Although click-through logs can provide implicit feedback of users＇ click preferences, de- riving accurate absolute relevance judgments is difficult be- cause of the existence of click noises and behavior biases. Previous studies showed that user clicking behaviors are bi- ased toward many aspects such as ＂position＂ （user＇s attention decreases from top to bottom） and ＂trust＂ （Web site reputa- tions will affect user＇s judgment）. To address these problems, researchers have proposed several behavior models （usually referred to as click models） to describe users？ practical browsing behaviors and to obtain an unbiased estimation of result relevance. In this study, we review recent efforts to construct click models for better search ranking and propose a novel convolutional neural network architecture for build- ing click models. Compared to traditional click models, our model not only considers user behavior assumptions as input signals but also uses the content and context information of search engine result pages. In addition, our model uses pa- rameters from traditional click models to restrict the meaning of some outputs in our model＇s hidden layer. Experimental results show that the proposed model can achieve consider- able improvement over state-of-the-art click models based on the evaluation metric of click perplexity.
