Data-Driven Similarity Measures for Matrimonial Application

Chia, Yong Fang (2020) Data-Driven Similarity Measures for Matrimonial Application. Final Year Project, UTAR.

Preview

PDF Download (19Mb) | Preview

Abstract

Marriage is a life-long commitment that completes our life. It will widen our horizons and the meaning of life on this Earth. However, the marriage rate in Malaysia continues to decline. Late marriage is one of the reasons that led to the decline of the marriage rate. Many people tend to get married later because of the difficulties in seeking a suitable spouse. Offline dating is time-consuming and limited by geographic proximity. Due to the convenience provided by the Internet, online dating has become a new trend in seeking potential partners. Hence, this project aims to develop a web-based matrimonial application that enables people to find their potential partner for marriage. Five similarity measures were proposed in this project to overcome the limitations of rule-based approach and Standard Query Language (SQL). The five similarity measures included Jaccard Coefficient, Cosine Similarity, Euclidean Distance, Manhattan Distance and Minkowski Distance. The application used the similarity measures to perform matching based on user preferences. In this project, the adopted software development methodology was phased development, which divided the development process into several phases. After the completion of system implementation, remote usability testing was conducted to evaluate which similarity measure is effective in finding matches that suit user preferences. The sample user data used for testing were collected from 85 people through a questionnaire. The test results showed that the match result obtained by Manhattan Distance was better then the other similarity measures. At the end of the project, all the objectives had been achieved. People can use the application to find their potential matches for marriage as per their priorities

Actions (login required)