Presentation
Estimating Housing Prices in Florida via Machine Learning
SessionPoster Session
Authors
Event Type
Posters
Receptions
Student Posters
VisDataAnalytics
Applications
Student Program
Facilitation
Workforce
TimeTuesday, July 246:30pm - 8:30pm
LocationKings Garden 3-4-5
DescriptionA homebuyer would always want to get the best deal, house at the lowest price. The problem of determining the best possible price for a house is very complicated. The only way to determine the proper price would be look at data of the various factors that determine the house demand, and the real estate price trend.
The project will involve study of the real estate market in Florida, build a model for price estimation of residential homes, by looking at the relationships between 68 different factors that will determine the house price, including frequency of various natural disasters that will impact the market demand for a certain period of time . The data will be from Zillow website .Data for the first 2000 different types of homes, classified based on primary factors, will be used and a model will be developed using Decision Trees and Neural network, and also clustering method. The methods will be implemented using Python, to identify the trends in different areas, as the data is available. Once a model based on the relationships is identified, it will be tested to predict the prices of any random house in the training data. Based on the errors values modifications can be made to the models, which will be later used to predict prices for a different test set, and the process will be continued till we have a robust model.
This poster will give information of the data used, tools applied for analysis and the model developed.
The project will involve study of the real estate market in Florida, build a model for price estimation of residential homes, by looking at the relationships between 68 different factors that will determine the house price, including frequency of various natural disasters that will impact the market demand for a certain period of time . The data will be from Zillow website .Data for the first 2000 different types of homes, classified based on primary factors, will be used and a model will be developed using Decision Trees and Neural network, and also clustering method. The methods will be implemented using Python, to identify the trends in different areas, as the data is available. Once a model based on the relationships is identified, it will be tested to predict the prices of any random house in the training data. Based on the errors values modifications can be made to the models, which will be later used to predict prices for a different test set, and the process will be continued till we have a robust model.
This poster will give information of the data used, tools applied for analysis and the model developed.