This paper discusses query expansion techniques in information retrieval, focusing on the comparison between global and local analysis methods. The authors argue that while global analysis, which examines the entire corpus to discover word relationships, has some advantages, local analysis—based on the documents retrieved by the initial query—is generally more effective. They also propose a new technique called local context analysis, which combines elements of both global and local methods. The study evaluates these techniques using different query sets and corpora, such as TREC and WEST. The results show that local context analysis outperforms both global analysis and simple local feedback in terms of retrieval effectiveness and predictability. The paper also highlights the importance of parameter tuning and the impact of the number of passages or documents used for expansion. Overall, the findings suggest that local analysis, particularly local context analysis, is a more effective approach for query expansion in information retrieval systems.This paper discusses query expansion techniques in information retrieval, focusing on the comparison between global and local analysis methods. The authors argue that while global analysis, which examines the entire corpus to discover word relationships, has some advantages, local analysis—based on the documents retrieved by the initial query—is generally more effective. They also propose a new technique called local context analysis, which combines elements of both global and local methods. The study evaluates these techniques using different query sets and corpora, such as TREC and WEST. The results show that local context analysis outperforms both global analysis and simple local feedback in terms of retrieval effectiveness and predictability. The paper also highlights the importance of parameter tuning and the impact of the number of passages or documents used for expansion. Overall, the findings suggest that local analysis, particularly local context analysis, is a more effective approach for query expansion in information retrieval systems.