Experimental data set preparation and management application for information retrieval
Project no. 176 · Status: Kapanmış · 2018
Summary
The academic studies conducted in the area of Information Retrieval (IR) should be supported by experimental findings. For this purpose, it is important to establish the data sets suitable for the IR process and to evaluate the performance correctly. Specifically, the process of evaluating whether the documents conform to the query objectives is rather complex and time consuming. In another aspect, the differences in the IR requirement shape both the data set and the performance criteria. In this project, a web-based application has been developed, which is a demanding process, and which prepares data sets that are appropriate for a specific task and selects the appropriate criteria and calculates the performance of the systems. TREC standards were used during the preparation of the dataset - documents, queries and relevant documents. Performance evaluation criteria are based on the snapshot-query task and a structure is created in which different criteria can be added easily.