Recommendation System in E-house Mobile Application
Tran Thanh Loc Nguyen · UTPedia (Universiti Teknologi Petronas) · 2020
In recent years, real estate domain provided huge amount of data and information of property on the Internet. User is getting trouble because they spend lot of time and efforts to search for good properties match their need, then contact with several peoples and set up meeting to give final decision. The solution of this has led to the development of recommender system, which is a powerful technique that apply knowledge discovery and understand user behavior based on user data/rating on past-experience to solve the current problem, generate recommended property that suit user taste. This Final Year Project 2 report aim to develop a Recommender System that can explore and automated suggest property to user. Hence, literature review discussed about the importance and efficiency of Recommendation System in modern world and methodology demonstrate an implementation of Recommender based on property data sets that author collected from multiple users by learning their preference, process it and suggest list of properties that most suitable. The system is developed by using Surprise Python package framework. Throughout the results of this project, author compared between two most common method for the establishing of Recommender that are top�N and kNN. Each method has advantages and disadvantages but overall top�N give better performance and fit within real estate property system model. Recommender System represent a good basis need to property market and bring solutions to many problems for better application and enhance quality of life.