Abstract
With the improvement of residents' consumption level, consumers' demand for all kinds of imported consumer goods has gradually increased, but its quality problem can not be ignored. The quality governance of consumer goods is particularly important, which is a part of social co-governance. It emphasizes the joint participation and responsibility to quality of multiple-subject, such as media, government, enterprises, social groups and the public. In the Internet age, social media is playing an increasingly important role in the reporting of various major event relative to commodity quality. Therefore, the research on the quality attention of imported consumer goods based on text mining to online news can be regarded as a valuable exploration. In this paper, using 3 identified standards —— news authority, news crawling volume and website traffic, we select 2 news platforms, i.e., China Quality News Network as well as Global Textile Network which have the most synthesis advantages to be qualified publishing quality news of imported consumer goods. Then, text mining on online news related to imported garments quality has been conducted by Python through the following efforts: (1) Garments’ quality dictionaries including quality attribution and quality scoring, has been developed by ourselves. (2) This paper verifies the rationality of the dictionary by using a Radial Basis Function (RBF) in Support Vector Machine (SVM) algorithm. The average accuracy on the training data set of the quality attribution dictionary is as high as 93.49%, showing that the selective attribution indexes in the quality attribution dictionary is rational. (3) From the above-mentioned news platforms, all the news relative to garments’ quality until September 10, 2020, has been crawled. There are 10780 crawled news totally, in which 8534 are effective news selected by removing duplicate and other data preprocessing methods. (4) On the basis of Maslow's Hierarchy of Needs, the 33 quality attributions in the dictionary have been divided into 4 categories: Physiological need attribution, Safety attribution, Functional attribution, Cognitive & aesthetic attribution. Text analysis has been made to the media attention about quality of imported consumer goods from the 4 categories, and time series analysis has also been made. (5) The media quality score of quality attribution has been calculated and analysed.)6) A correlation analysis of the media attention and score of each attribute category has been conducted.
The results come to the following conclusions: (1) Different platforms basically share the same concern on the attribution quality of imported garments. The top five quality attributions with high media attention are harmful substances, typical mechanical & physical properties, logo, aesthetic, and comfort. The above 5 quality attributions mainly belong to the safety attribution and cognitive & aesthetic attribution. (2) The higher quality score, the higher media attention, and the worse the quality of imported garments. The media attention on the quality of imported garments experienced four stages: the sustained level stage (2001-2003), the slow growth stage(2003-2012), the rapid growth stage (2012-2014)and sustained pullback stage(2014-2019). (3) Media attention has high correlation to the quality of quality score of imported garments. (4) From 2014 to the present, the attributions quality of imported garments from low to high are Safety attribution, Cognitive & aesthetic attribution, Functional attribution, Physiological need attribution, which lead to the media attention to the attributions from high to low are Safety attribution, Cognitive & aesthetic attribution, Functional attribution, Physiological need attribution. (5) During the period of the rapid growth stage (2012-2014), the totality quality of imported garments declined significantly, give rise to the media attention also increased significantly. (6) After 2014, the totality quality of imported garments showed a trend of gradual improvement, and then the media attention of imported clothing quality was also declining. (7) The quality of safety attribution changes the most, but it improves the fastest.
According to the analysis of this paper, media attention has a very close relationship with the quality score of imported garments. The more serious the quality problems, the more media attention would be drawn to them. The high media attention quality attributes need more attention of government to strengthen quality supervision. Finally, this paper puts forward some suggestions on the quality governance of imported consumer goods.