We present a statistical method called Covering Topic Score （CTS） to predict query performance for information retrieval. Estimation is based on how well the topic of a user＇s query is covered by documents retrieve...We present a statistical method called Covering Topic Score （CTS） to predict query performance for information retrieval. Estimation is based on how well the topic of a user＇s query is covered by documents retrieved from a certain retrieval system. Our approach is conceptually simple and intuitive, and can be easily extended to incorporate features beyond bag- of-words such as phrases and proximity of terms. Experiments demonstrate that CTS significantly correlates with query performance in a variety of TREC test collections, and in particular CTS gains more prediction power benefiting from features of phrases and proximity of terms. We compare CTS with previous state-of-the-art methods for query performance prediction including clarity score and robustness score. Our experimental results show that CTS consistently performs better than, or at least as well as, these other methods. In addition to its high effectiveness, CTS is also shown to have very low computational complexity, meaning that it can be practical for real applications.展开更多
Performance predictions for database queries allow service providers to determine what resources are needed to ensure their performance. Cost-based or rule-based approaches have been proposed to optimize database quer...Performance predictions for database queries allow service providers to determine what resources are needed to ensure their performance. Cost-based or rule-based approaches have been proposed to optimize database query execution plans. However, Virtual Machine （VM）-based database services have little or no sharing of resources or interactions between applications hosted on shared infrastructures. Neither providers nor users have the right combination of visibility/access/expertise to perform proper tuning and provisioning. This paper presents a performance prediction model for query execution time estimates based on the query complexity for various data sizes. The user query execution time is a combination of five basic operator complexities： O（1）, O（log（n））, O（n）, O（nlog（n））, and O（n2）. Moreover, tests indicate that not all queries are equally important for performance prediction. As such, this paper illustrates a performance-sensitive query locating process on three benchmarks： RUBiS, RUBBoS, and TPC-W. A key observation is that performance-sensitive queries are only a small proportion （20%） of the application query set. Evaluation of the performance model on the TPC-W benchmark shows that the query complexity in a real life scenario has an average prediction error rate of less than 10% which demonstrates the effectiveness of this predictive model.展开更多
基金the National Natural Science Foundation of China under Grant No.60603094the National Grand Fundamental Research 973 Program of China under Grant No.2004CB318109
文摘We present a statistical method called Covering Topic Score （CTS） to predict query performance for information retrieval. Estimation is based on how well the topic of a user＇s query is covered by documents retrieved from a certain retrieval system. Our approach is conceptually simple and intuitive, and can be easily extended to incorporate features beyond bag- of-words such as phrases and proximity of terms. Experiments demonstrate that CTS significantly correlates with query performance in a variety of TREC test collections, and in particular CTS gains more prediction power benefiting from features of phrases and proximity of terms. We compare CTS with previous state-of-the-art methods for query performance prediction including clarity score and robustness score. Our experimental results show that CTS consistently performs better than, or at least as well as, these other methods. In addition to its high effectiveness, CTS is also shown to have very low computational complexity, meaning that it can be practical for real applications.
文摘Performance predictions for database queries allow service providers to determine what resources are needed to ensure their performance. Cost-based or rule-based approaches have been proposed to optimize database query execution plans. However, Virtual Machine （VM）-based database services have little or no sharing of resources or interactions between applications hosted on shared infrastructures. Neither providers nor users have the right combination of visibility/access/expertise to perform proper tuning and provisioning. This paper presents a performance prediction model for query execution time estimates based on the query complexity for various data sizes. The user query execution time is a combination of five basic operator complexities： O（1）, O（log（n））, O（n）, O（nlog（n））, and O（n2）. Moreover, tests indicate that not all queries are equally important for performance prediction. As such, this paper illustrates a performance-sensitive query locating process on three benchmarks： RUBiS, RUBBoS, and TPC-W. A key observation is that performance-sensitive queries are only a small proportion （20%） of the application query set. Evaluation of the performance model on the TPC-W benchmark shows that the query complexity in a real life scenario has an average prediction error rate of less than 10% which demonstrates the effectiveness of this predictive model.
