Research On Product Sensual Image Vocabulary Based On Text Mining

With the development of the times,the demand for products from consumers has been transformed from simple functional requirements to emotional needs.How to accurately obtain user needs and design products that meet the emotional needs of users has become one of the important challenges faced by contemporary designers.In this era,sensibility engineering is increasingly favored by scholars as a discipline that pays great attention to user feelings and needs.The spread and application of data center has brought new opportunities to Perceptual engineering.However,in today's Perceptual engineering research,the application of data center technology is not extensive,and some studies are not deep enough due to lack of pertinence.In order to solve the problems,firstly this article reviews the status quo of literature research on perceptual image vocabulary,and points out that the traditional method of vocabulary extraction has high labor cost,too much dependence on experts and excessive interference factors.Based on the perceptual image theory and the networking comment mining in data center,this article proposes a method for extracting perceptual image vocabulary on word vector clustering.This method does not require questionnaire survey and expert intervention,and the algorithm completes the calculation statistics,thus greatly reducing the labor cost of the perceptual image vocabulary extraction experiment.And improves the problem that the traditional vocabulary extraction method cannot obtain a large number of user samples,and has good expandability.The method mainly includes the threes steps:1.Acquisition and establishment of the original corpus.Get product review text data through Python software.2.The establishment of a dimensional semantic network.Based on the high-frequency vocabulary,a dimensional semantic vocabulary network is established for the extraction experiment of perceptual image vocabulary.The purpose is to distinguish the evaluation dimension of the comment and extract the adjectives that are only for the appearance of the appearance,which is used to extract the sensuous imagery words.3.Extraction of perceptual imagery vocabulary.The method is based on the similarity of lexical word vectors with similar meanings of word vectors.The word2 vec model is used to output the original sensible vocabulary and its word vector,and the original sensible vocabulary and its word vector are output.The word vectors of the original perceptual image vocabulary are clustered,and representative perceptual image vocabulary is extracted according to the central point of the k-means cluster.Based on the acquisition of the word vector,this paper also carried out the component coding experiment.Based on the word vector,the user's attention hotspot is analyzed to encode the product parts.This article takes the gas stove as a research case,the neural network technology is used to establish the connection between the emotional imagery vocabulary and the product modeling,and the image modeling mapping system is constructed.It enables the designer to obtain a specific product styling recommendation by inputting the user image value,and more intuitively understand the user's emotional image of the product.Finally,make summarizes and points out the direction for further research.