| Summary: | 碩士 === 國立臺北科技大學 === 電機工程系 === 107 === In recent years, information technology has developed rapidly, and many commercial applications have been combined with big data information technology. This thesis focuses on the location selection of the store. We apply machine learning technology to analyze whether a location is suitable for site selection. However, this information requires the use of manpower to assess and increase costs. Therefore, in this thesis, we developed a platform that allows users to quickly and succinctly understand these external environment features as they choose their location.
In this thesis, we develop a site selection forecasting system. Based on collecting government public information and crawling the data of Taipei Government Cloud Platform and concatenating Google API data, the external environment features of each location are put into decision tree model to make the prediction. When the user enters an address on the system, the system immediately collect the external environment features near the address and transmit the data to the system backend for prediction, providing users with better information when selecting the site.
This system mainly uses the Python Flask framework to build the Web, develop with the MySQL database, and use Python for machine learning training and prediction. There are four models in the system, namely 711_store model, Family_store model, hilife_store and OK_store model. The accuracy of 711_store model test data reaches more than 70%, and the other three models test data reaches 60% accuracy rate.
|
|---|
