International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395 -0056
Volume: 04 Issue: 03 | Mar -2017
p-ISSN: 2395-0072
www.irjet.net
Real Estate Investment Advising Using Machine Learning Dr. Swapna Borde1, Aniket Rane2, Gautam Shende3, Sampath Shetty4 1Head
Of Department, Department Of Computer Engineering, VCET, Mumbai University, India Department Of Computer Engineering, VCET, Mumbai University, India 3Student, Department Of Computer Engineering, VCET, Mumbai University, India 4Student, Department Of Computer Engineering, VCET, Mumbai University, India 2Student,
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The project makes a comparative study of various Machine Learning algorithms namely Linear Regression using gradient descent, K nearest neighbor regression and Random forest regression for prediction of real estate price trends. The aim of this paper is to examine the feasibility of these machine learning algorithms and select the most feasible one. To achieve the aim, parameters like Living Area, Number of rooms, Distance from airport/highway/station/major landmarks, Proximity to hospitals, Shopping options, Number of theaters, geographical location(harbor/central/western) are used as the input to the model and real estate price in the next quarters is the output variable. The quarterly data during 2005-2016 is employed as the data set to construct the model and the data has been obtained using Web Scraping from websites like 99acres.com, Magicbricks.com, Google.com. The experimental results on the training data set are used to compare the various algorithms based on error calculation using Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MAE).
Key Words: Linear Regression, K Nearest Neighbor (KNN), Random Forest algorithm, Decision Tree, forecast
1. INTRODUCTION Prices of real estate properties are critically linked with our economy. Accurately predicting real estate prices is not possible. However, prediction of real estate trends is a realistic prospect. Yet, we do not have accurate measures of housing price trends based on the vast amount of data available. In Mumbai alone there are around 10,000 current listings of 350 areas or more at 99acres.com. This rich dataset should be sufficient to establish a regression model to accurately predict the real estate prices in Mumbai. A property’s appraised value is important in many real estate transactions such as sales, loans, and its marketability. Traditionally, estimates of property prices are often determined by professional appraisers. The disadvantage of this method is that the appraiser is likely to be biased due to vested interest from the lender, mortgage broker, buyer and seller. Therefore, an automated prediction system can serve
For the buyers of real estate properties, an automated price prediction system can be useful to find under/overpriced properties currently on the market. This can be useful for first time buyers with relatively little experience, and suggest purchasing strategies for buying properties. This paper makes a comparative study of the three mentioned algorithms viz. Linear Regression using gradient descent, KNN Regression, Random forest regression for analyzing the trends. The data set extracted is split into training set and testing set in a ratio of 80:20 or 4:1. Finally, based on the conclusions drawn from the comparative study of the algorithms are used to develop a front end that suggests the user areas that would be most profitable for investment. On the front end, the user is asked for parameters such as Total budget and Total Area and based on the prediction of the prediction model which implements the most feasible algorithm for our data set, the user is provided with suggestions for investment
2. PREDICTION BASED ON LINEAR REGRESSION USING GRADIENT DESCENT Linear Regression principle is basically used for prediction and forecasting. Being the simpler and basic technique as compared to other machine learning algorithms, the implementation is less complex but the degree of error is slightly above the average. In this paper, we focus on Liner Regression using Gradient Decent for regression and forecast. Linear Regression using Gradient Descent is a learning machine which estimates a function according to a given data set G={(xi, yi)}n, where xi denotes the input vector, yi denotes the output value and n denotes the total number of the data. If the learning machine is working on multiparameters, xi denotes the input matrix and yi denotes the output vector.
as an important third party source which is not biased. Š 2017, IRJET
|
Impact Factor value: 5.181
|
ISO 9001:2008 Certified Journal
|
Page 1821