Research On Inventory Control And Reprint Strategy Of Textbook Based On Data Mining

In recent years,Chinese book publishing industry has made considerable progress.The scale and profit of book publishing have been steadily increasing.Under the excellent situation,there is a problem of faster inventory accumulation.Exploring the status quo of the publishing industry,on the one hand,textbooks,as an important part of book publishing,account for nearly 40% of the total publishing scale and retain the important position of the publishing company's major profit sources.On the other hand,from the perspective of the new and reprinted forms of publication,the scale of reprint is far higher than the new edition and still maintains rapid growth.How to develop a scientific and reasonable reprint strategy for the publishing of teaching materials is very urgent and important for the publishing industry.The current research is more about empirical qualitative analysis of reprint strategies or modeling from the perspective of economic production.However,for the actual operation of the publishing industry,it is often necessary to first meet the market's product demand,and consider the cost factors such as batch and time to make decisions based on this.Therefore,it is of great significance to make forward-looking predictions on the production of teaching materials and then guide the production of enterprises.This article introduces the current publication of textbooks in the publishing industry,explains the concepts of reprinting of textbooks and the problems existing in current operations.With the accumulation of time,publishing companies will inevitably accumulate a large amount of data related to the reprint of teaching materials in production,including production,sales,inventory and other aspects.Taking into account the characteristics of data and analysis objectives,big data mining technology can make full use of these data to find out the influencing factors and internal laws of reprinting of teaching materials,and provide reference for reprint production.This paper selects a combination forecasting method——random forest algorithm to establish the reprint model of the teaching materials: analyzing the relevant influencing factors of the reprint of the teaching materials and establishing a characteristic system,according to the two-step strategy from the reprinting mode selection to the teaching material sales forecasting,respectively establishing the corresponding Classification model and forecasting model,and made multi-angle processing of direct sales forecasting and discretized forecasting for sales forecasting model.In the empirical stage,the article starts with the real data of a publisher's ERP system,collects the materials such as the purchase and sales of materials and sales details as the original data,and selects the 1994 reprint samples as the model's research object.In the reprinting classification problem,the model established by the random forest algorithm has a higher accuracy than the classical logistic regression model.In the sales forecast,the evaluation scores predicted by the C4.5 algorithm and the random forest algorithm are compared,and random forests achieve a lower forecasting bias.Considering the batch problem in actual production,we choose two ways to predict the sales volume.The result shows that the discretization prediction has better practical guidance value.Finally,the practical value and effect of the model have been recognized by a certain publisher and applied to the reprinting method selection and sales forecast of 2018 textbooks production.Some calculation results are shown in the text.