2023 ADM+S Symposium Poster Competition
Nuha Abu Onq, RMIT University
Title: User Query Variations and Cognitive Complexity of
Search Tasks
When given a common information need, a large number of searchers tend to form a wide variety of queries. However, one aspect of this query variation that remains unexplored is the extent to which searchers rely on their past experience when forming queries. Our research focuses on this aspect and also considers the complexity of information needs and its connection to query formulation. In this study, we categorise each term in users' initial queries to examine its relationship with the task description. To the best of our knowledge, this level of query term analysis has not been previously explored. By examining the terms used by searchers, we delve into the realm of query variants that has received limited exploration. Our contributions include: (1) an examination of the terms used in variants and backstories; (2) an understanding of those terms and their relationship to a set of existing backstories newly categorised into cognitive complexity.
Our investigation revealed an interesting pattern in users’ query formulation process. Understanding the methods employed by individuals in generating queries, including the types and frequency of novel terms used, can provide valuable insights for researchers. This knowledge can aid in the development of more effective models for generating automated query variants that closely resemble those created by users. By incorporating these user-centric approaches, researchers can enhance the relevance and accuracy of automated query-generation techniques. Moreover, drawing upon the implications of information retrieval (IR) design, prior research has recommended the integration of comprehensive query autocompletion to enhance search queries. This approach involves considering the quality and ranking of current suggestions and identifying potential enhancements based on user preferences [1]. The findings of our study shed light on users’ initial inclination toward general information, such as data from Wiki, but their subsequent shift toward more specialised aspects as they refine their queries [2]. By incorporating these insights, search engines can improve the quality and ranking of query autocompletion suggestions. This can be achieved by initially presenting broad suggestions and gradually transitioning to more specialised ones. Furthermore, the integration of contextual based, demographic-based, time-sensitive, and location elements in query autocompletion can further enhance its relevance and quality, providing tailored suggestions aligned with users’ context, demographics, and temporal relevance. Additionally, individuals may exhibit inherent bias within their search queries, prompting the need for a comprehensive understanding of query variation. Thus, it becomes imperative for individuals to recognize and acknowledge this particular aspect.
[1] Fei Cai and Maarten de Rijke. 2016. A Survey of Query Auto Completion in Information Retrieval. Found. Trends Inf. Retr. 10, 4 (2016), 273–363.
[2] Jia Chen, Jiaxin Mao, Yiqun Liu, Fan Zhang, Min Zhang, and Shaoping Ma. 2021. Towards a Better Understanding of Query Reformulation Behavior in Web Search. In Proc. WWW. 743–755.
2023 ADM+S Symposium Poster Competition
Nuha Abu Onq, RMIT University
Title: User Query Variations and Cognitive Complexity of
Search Tasks
When given a common information need, a large number of searchers tend to form a wide variety of queries. However, one aspect of this query variation that remains unexplored is the extent to which searchers rely on their past experience when forming queries. Our research focuses on this aspect and also considers the complexity of information needs and its connection to query formulation. In this study, we categorise each term in users' initial queries to examine its relationship with the task description. To the best of our knowledge, this level of query term analysis has not been previously explored. By examining the terms used by searchers, we delve into the realm of query variants that has received limited exploration. Our contributions include: (1) an examination of the terms used in variants and backstories; (2) an understanding of those terms and their relationship to a set of existing backstories newly categorised into cognitive complexity.
Our investigation revealed an interesting pattern in users’ query formulation process. Understanding the methods employed by individuals in generating queries, including the types and frequency of novel terms used, can provide valuable insights for researchers. This knowledge can aid in the development of more effective models for generating automated query variants that closely resemble those created by users. By incorporating these user-centric approaches, researchers can enhance the relevance and accuracy of automated query-generation techniques. Moreover, drawing upon the implications of information retrieval (IR) design, prior research has recommended the integration of comprehensive query autocompletion to enhance search queries. This approach involves considering the quality and ranking of current suggestions and identifying potential enhancements based on user preferences [1]. The findings of our study shed light on users’ initial inclination toward general information, such as data from Wiki, but their subsequent shift toward more specialised aspects as they refine their queries [2]. By incorporating these insights, search engines can improve the quality and ranking of query autocompletion suggestions. This can be achieved by initially presenting broad suggestions and gradually transitioning to more specialised ones. Furthermore, the integration of contextual based, demographic-based, time-sensitive, and location elements in query autocompletion can further enhance its relevance and quality, providing tailored suggestions aligned with users’ context, demographics, and temporal relevance. Additionally, individuals may exhibit inherent bias within their search queries, prompting the need for a comprehensive understanding of query variation. Thus, it becomes imperative for individuals to recognize and acknowledge this particular aspect.
[1] Fei Cai and Maarten de Rijke. 2016. A Survey of Query Auto Completion in Information Retrieval. Found. Trends Inf. Retr. 10, 4 (2016), 273–363.
[2] Jia Chen, Jiaxin Mao, Yiqun Liu, Fan Zhang, Min Zhang, and Shaoping Ma. 2021. Towards a Better Understanding of Query Reformulation Behavior in Web Search. In Proc. WWW. 743–755.