Fashion Clothing Design Style Recommendation using Extra Tree Classifier Combined with LSTM

Recommendation System plays an vital part in e-commerce requisition as they help the consumer or user to choose the requires items within limited time. Presently, several e-commerce use a text-based artifact search which has problem to fetch the most alike products of the system. However, most cloth recommendation have several limitations such as; high source consumption, inconsistence subjectively cloth labeled. To overcome above limitation this paper is mainly focused to Extra Tree Classifier (ETC) based feature sub categorization method to overcome the high time consuming and provide intelligent recommendation of clothing style. Convolution Neural Network (CNN) used to recognize effective features for future recommendation. To train the identified aspects with to make efficient rating prediction the numerical function of LSTM based feature extraction are used more effective enables user to select the most rated item (clothing) of the function. The proposed ETC-LSTM method increases exploration and exploitation which helps to improve the user recommendation based rating. The production of the ETC-LSTM method in analyzed in the terms of Area Under Cover (AUC), Precision and Recall. The ETC-LSTM method has 95.347% than the existing MAlexNet-JSEPCC, DVESM-CAE and SLT-DT.