<![CDATA[As a data provider, BPS provides services for the public to access BPS data. One of these services is the search service on the BPS website. However, the search services have not met consumer expectations. One of the efforts that can be done to meet consumer expectations is to increase search effectiveness to make it more relevant to the user's intent. Therefore, this study aims to build a query classification function on a search engine and test whether this function can improve search effectiveness. The query classification function is built using a machine learning model. We compare five algorithms, those are SVM, random forest, Gradient Boosting, KNN, dan Naive Bayes. Of the five algorithms used, the best model is obtained from the SVM algorithm. Then, the function is implemented on a search engine whose effectiveness is measured based on precision and recall values. As a result, the query classification function can narrow the search results to a particular query, thereby increasing the precision value. However, the query classification function does not affect the recall value.]]>
Copyrights © 2021
