Shopping Mall Brand Store Efficiency Classification Forecast Based On Co-Forest Algorithm

With the development of the economy,shopping centers have sprung up everywhere in our country,followed by fierce competition between shopping centers.The slightest carelessness in each part of the shopping mall construction will lead to the failure of the construction of the entire shopping center,resulting in huge losses.Choosing brand is very important for a shopping center,and usually determines whether the shopping center and the brand can achieve a win-win situation.Whether a brand is suitable for settling in a shopping mall,the brand's store efficiency is the most intuitive method of judgment,but settle in a brand to obtain its efficiency data is time consuming and costly for a shopping center,using machine learning methods to use the efficiency data of brand settled in the target shopping center and the feature data of all the brand to make significant predictions is very important.First of all,it analyzes the influencing factors of the brand store's efficiency,collect the features that can describe the brand,and obtain the affect to brand's store efficiency of this features.Get data of this features,mainly get data of five aspects,include brand online sales data,brand awareness data,brand popularity-related data,brand's own attributes,and other shopping malls' brand data.Secondly,with regard to the introduction of brands in other domestic shopping malls except for the target mall,use bipartite graph to calculate the data obtained from other malls to introduce brands,then the recommendation introduction index for each brand introduced by other shopping centers was obtained,and this was taken as the one-dimensional feature describing the brand,and used as the input of Co-Forest along with other features describing the brand.Again,the brand recommendation index calculated from the bipartite graph will be integrated with the acquired data of other brand-related characteristics,and the efficiency of the brand that has been introduced into the target shopping center will be the labeled data,and the brand not introduced into the target shopping center will be used as the unlabeled data.Use the Co-Forest algorithm to combine labeled and unlabeled data to predict each brand's store efficiency classification.Finally,taking the Aegean Shopping Park of Hongxing Commercial as an example,use the efficiency data of the brand which has beed settled in it,the proposed method has been experimentally verified.Compare the experimental results with the ordinary supervised learning algorithm and analyze the pros and cons of each algorithm.