Research On Evaluation Algorithm Of Logistics Service Quality Based On Sentiment Analysis

Text sentiment analysis is an important branch of natural language processing,but there are still some shortcomings in the research of sentiment analysis algorithm.The sentiment analysis algorithm based on traditional machine learning needs to train a large amount of data,which leads to its slow execution rate and a lot of redundant work,so deep learning algorithm is proposed.However,a single deep learning algorithm is still unable to solve the emotional polarity problems of polysemy,synonyms and synonyms.To solve the above problems,this paper proposes an improved sentiment analysis algorithm based on multi-layer cooperative convolutional neural network LSTM-TTT,which solves the problems of low model training efficiency and large amount of redundant data of traditional sentiment analysis algorithms.At the same time,compared with a single deep learning algorithm,the sentiment analysis algorithm based on LSTM-TTT can fully extract contextual semantic information,and accurately identify the synonyms,synonyms and other easily confused emotion polar words in the text.The main work of this paper focuses on the following aspects:(1)Study the principle of emotion analysis algorithm based on deep learning,explore and analyze Transformer,LSTM,Text-CNN and TF-IDF algorithms,and integrate LSTM-TTT algorithm based on emotion analysis through collaborative learning and dynamic full connection of multi-layer neural network.And design LSTM-TTT algorithm parameter training strategy comparison experiment to train the algorithm performance index to the optimal state.(2)Put forward the text data set of Tmall goods online logistics service quality evaluation,obtained the logistics service quality evaluation data of different types of goods in Tmall supermarket by crawler technology,and divided the training set,verification set and test set according to proportion.At the same time,the paper analyzes the stylistic characteristics of the evaluation of logistics service quality on the e-commerce platform,and carries on the text preprocessing.(3)Design a comparison test between LSTM-TTT algorithm and four machine learning algorithms,namely Random Forest,XGBoost,LSTM and Transformer.The experimental results show that the accuracy and F1 value of LSTM-TTT algorithm after training is 95.5%,which has achieved excellent results.Compared with LSTM algorithm,the accuracy is improved by 6%.Compared with Transformer algorithm,the accuracy is increased by 8%,which proves the effectiveness and feasibility of LSTM-TTT algorithm.LSTM-TTT sentiment analysis algorithm is an improved deep learning algorithm,which has enriched the relevant research of sentiment analysis in natural language processing.At the same time,LSTM-TTT sentiment analysis algorithm is applied to the text data set of online logistics service quality evaluation for training analysis,which is an application innovation in the field of logistics service quality monitoring in e-commerce industry.