Research And Development Of Material Purchase Follow Up System For Customized Equipment Enterprises

For customized equipment enterprises,it is a unique competitiveness and a great challenge to meet the customized needs of customers.The randomness of order quantity and the continuous change of product requirements bring great uncertainty to the project,and the pressure of delivery time also puts forward higher requirements for enterprise operation management.For small and medium-sized enterprises,manufacturing orders are usually prepared according to the quantity and quantity of the equipment,which are not specified in the purchase order In the past,material purchase orders are often delayed.In this paper,we use supervised machine learning to predict the tardiness of the purchase order from the perspective of material tardiness.At the same time,we use Delphi method to build a follow-up strategy for orders with different characteristics.First of all,through the analysis of the characteristics of customized equipment enterprises,combined with the specific procurement business process of an enterprise,the factors that may affect the material procurement progress are analyzed in detail,and they are divided into five categories.After preprocessing and analyzing the enterprise data,the original features of the data are extracted according to the five categories of factors.The features are constructed by feature engineering,the number of features is expanded,and the feature selection method of ensemble learning is used to select features.On this basis,six kinds of supervised learning classifiers are constructed.Experiments are designed and calculated to compare the effects of different feature selection methods and classifiers,and the most effective tardiness prediction model is obtained.At the same time,several main features which have great influence on the model decision are found out.Secondly,in order to construct the material purchase order tracking strategy,the main features extracted from the supervised learning model were used to classify the purchase orders.After data collection and discussion,a questionnaire was designed for the construction of material purchase order tracking strategy,and 18 experts in related fields were selected for investigation and consultation.After three rounds of repeated expert consultation,the experts’ opinions tend to be consistent.Based on the results of expert consultation,the initial strategy of all kinds of purchase order tracking is basically determined.In addition,a set of real-time dynamic adjustment rules is designed according to the actual situation,trying to achieve better effect of follow-up.Finally,an app system of material purchase order tracking for customized equipment enterprises is designed and developed.The system embeds the supervised learning prediction model and the purchasing follow-up strategy for different orders,which can realize the tardiness prediction of material purchase order and the formulation and implementation of follow-up plan,and can improve the tardiness of material purchase to a certain extent.