Research On Application Of Publishing Topic Prediction Based On Book Sales Data

In recent years,extensive research and application of data mining technology has aroused great concern from all walks of life,all walks of life are urgently needed to convert enterprise data into useful information and knowledge.Book publishing industry develops in the challenge and opportunity,the Internet is hitting all walks of life,including the publishing industry,and at the same time bringing new opportunities for the publishing industry,the publishing industry will continue to integrate new computer technology,and finally give the birth of a new industrial model and industrial chain.It is imperative for the book publishing industry to use advanced data mining technology.In view of the problem that the published planning topics rely on the subjective experience and the potential problems of user needs and the current social hot research is not accurate,through the study of book market analysis,according to the short-term volatility of the book sales market and the long-term and periodic characteristics of the book sales market,this paper presents an integrated time series prediction algorithm and neural network algorithm respectively in the short-term and long-term topic prediction.In the medium and long-term forecast,the Holt-Winters time series forecasting model predicts book sales according to the book category to provide the basis for planning for the publishing unit to make a reasonable selection of topics as well as the time-consuming forecast.In doing so,the publishing houses can avoid huge losses caused by delayed sales season,and reduce unnecessary consumption of people,wealth and material in the process of book publishing.In the short-term forecast,the neural network model is used to forecast the printing quantity of the designated books in each region,and the prediction accuracy of the hot events is improved by the improved model of the author's heat weighting.The JSOUP framework used to crawl Sina micro-blog hot search author and hot search words,and build hot search information database.After the author's information and content information is judged by hot spot,the hot author content and popular content are weighted.In doing so,we can more accurately grasp the huge benefits of the "Nobel effect",and assist the book selection staff to make the correct judgment in the printing quantity arrangement.In this paper,we design a public account based on the book sales data to forecast the public account of the WeChat,and provide reliable selection of category and printing forecast for publishing staff.It can effectively control the market rule and meet the user's consumption tendency through the topic selection,it also can effectively reduce the opportunity to miss the best selling times and uneven distribution of printing,and then avoid the backlog of inventory and the consumption of a lot of manpower,material and financial resources,and effectively improve the economic efficiency of publishing units.