Research On Mobile Phone Review Based On LSTM Model

China's online shopping market is developing rapidly,and the transformation of consumption patterns has gradually been accepted by the public.A large number of online malls are booming,which brings convenience to people's lives and generates a lot of data in operation.The information comes from the description and introduction of the merchant,and the product reviews are mainly from real consumers with consumption records.From this point of view,the product text reviews not only help consumers understand the product status but also help Merchants upgrade products.In the era of big data,traditional statistical surveys are no longer sufficient for surveys of large amounts of text.Not only are they timeconsuming and labor-intensive,but the survey scope is small and the survey accuracy is not high.This situation is greatly unable to meet the needs of enterprises and individuals.Today,with the rapid development of artificial intelligence,some traditional investigation techniques are gradually being replaced.The relatively simple and convenient text mining related technologies to be used here.This article takes the reviews of the three major ecommerce companies—Taobao,Tmall,and JD.com 's children 's phone watches as examples.Based on web crawlers and natural language processing technology,it focuses on the characteristics of consumer demand for children 's mobile phones A series of statistical analysis and mathematical modeling are carried out on the comment text of the real phone watch.The following steps are mainly carried out: first,the original data of the model is collected,the collected data is applied to traditional statistical methods and machine learning algorithms such as word segmentation and word removal,and then the data is preprocessed using Python software Input into the constructed long and short-term memory neural network(LSTM)model to obtain text corpus with positive and negative emotional tendencies,which initially shows the advantages and disadvantages of phone watches.Finally,the LDA theme model was used to obtain the key themes of positive and negative sentiment texts.Finally,the demand for phone watches was found in the positive corpus,and the deficiencies of the phone watch were found in the negative corpus.After preprocessing and analysis of the LSTM model and LDA model,the consumer needs of special mobile phones for children and teenagers such as phone watches were extracted.These extraction results also provide product development for phone watch manufacturers and operators.The direction also provides a certain basis for the seller to formulate a marketing strategy.