Study On Location Of Chain Restaurants Based On Machine Learning Algorithm

Site selection,the first step of a restaurant's operation,determines the geographical location,business atmosphere,rent cost and customer composition of the restaurant,which plays a crucial role in the success of the restaurant.If the location decision fails,it is obvious that the restaurant will suffer a huge loss by the high rent costs and low profits.In the past time,restaurants mainly made location decision by their qualitative experience.However,such a way has a large site failure rate.Luckily,with the development of internet and artificial intelligence,we can make the site selection decision by big data technology today.This paper gains the data from e-commerce and map service platform based on the web crawler technology,applying machine learning algorithms to predict the suitable location of craft beer restaurants in Beijing.This paper meanly performed the following work.Firstly,we construct the hierarchical structure model of influencing factors on location selection problems,based on the theories related to retail stores site selection.Then,we find that the most important factors can be included as population factors,competition intensity,transportation convenience and other factors.Secondly,we crawl the data from DianPing,Lianjia,BaiduMap,and explore this original set of data combined with the feature of craft beer restaurants in Beijing,construct the training features and select it by the method of filtering and wrapping.Thirdly,we search the classification algorithm from the machine learning algorithm commonly used,such as GBDT,SVM,Random Forest,KNN and so on.Then we train the data set on those algorithm.Then,we select the appropriate figures to assess different algorithm,according to the result,we could find out the best classifier of this problem.Lastly,according to the results,we propose some suggestion to the location of craft beer restaurants.