Functional Attention Analysis Based On Deep Learning In Different Usage Scenarios Of Headphones

As a world-renowned electronic product manufacturer,Sony has always had a place in the field of earphone manufacturing.In recent years,the marketing point of earphone products has gradually shifted from the incomprehensible functional attribute parameters to the applicable scenarios of the earphones,and the use scenarios of the existing product positioning are mostly based on personal experience or questionnaires that lack a reasonable sampling plan,so it is concluded that the usage scenarios lack scientificity and credibility.This article takes the comments on all kinds of earphones in JD.com SONY's self-operated official flagship store as the research object,and conducts research on the focus of headset usage scenarios to make up for the lack of functional attributes that consumers care about from the perspective of usage scenarios.The following three issues are mainly studied.The first part is to get the main usage scenarios for consumers to use headphones.Firstly this paper uses Perplexity and Topic Coherence to determine the number of topics,and LDA topic model is established in the reviews with topics of 6.Then we summarize the topic results to obtain two types of topics:product usage scenarios and usage attitudes.The following product usage scenarios topics are summarized into four categories:transportation,sports and fitness,work and study,and daily entertainment.The second part is to train the classification model of the usage scenario.Manually annotating 50%of the comment data sets to train Word2vec+TextCNN and Bert's usescenario binary classification model,and found that the Bert model has a faster loss convergence speed and higher accuracy on the validation set.In the test set,the F1 score of the model under the scenarios of transportation,sports and fitness,leisure life,work and study is 77.1%,92.2%,78.6%,78.0%,and the precision is 79.4%,88.7%,74.1%,88.5%,and finally the number of comments under the corresponding scenarios is 745,783,959,and 653.The third part is to use dependency syntax analysis to extract the triples in the comments,and analyze the user concerns under different types of headphones and usage scenarios.It is found that the top three user concerns of different types of headphones have little difference,focusing on sound quality,comfort and noise reduction.In different scenarios,the user's concerns of different headset types are quite different,and the ranking of the same concerns is also different,as shown in Table 5.4.Based on the analysis results,this paper puts forward the following suggestions to manufacturers:(1)Different types of earphones have different usage scenarios,and focus on core user needs;(2)The same type of earphones need to focus on different functions to design multiple earphones,and pay attention to price positioning;(3)The public generally pays attention to the comfort and noise reduction function of headphones,so manufacturers should strengthens the design of ergonomic functions.