An Information Retrieval Approach to Finding Similar Questions in Question-Answering of Indonesian Government e-Procurement Services using TF*IDF and LSI Model

Abstract

In the implementation of e-procurement application in the government of the Republic of Indonesia, one of the important issues is how to find relevant questions in the archive of a question answering (QA) service with the question asked by the user. A common method used for finding relevant documents is representing text documents into vector space model (VSM). Relevance between query and documents can be calculated using document similarities theory, by comparing the deviation of angles between each document vector and the query (from user question) vector where the query is represented as the same kind of vector as the documents. The Vector Space model algorithms widely used are Term Frequency * Inverse Document Frequency (TF*IDF) and Latent Semantic Indexing (LSI), however both models have their respective limitation. Considering that problem, this paper proposed hybrid model that combines TF*IDF and LSI to fix some limitations on both. From the experimental results, it is found that the proposed model is outperform (P@1=0.67) compared to TF*IDF model (P@1=0.27) and LSI model (P@1=0.4) that stand alone.