A Method Of Material Procurement Progress Monitoring Based On Machine Learning

In the customized equipment industry,the uncertainty of order quantity and the shortened order delivery time brought by individual customization put forward higher pursuit of enterprise production management,and produce personalized at low cost and high efficiency in a shorter period.Order products are the unremitting pursuit of equipment manufacturing companies.The uncertainty of the order and its quantity will lead to the uncontrollable progress of material preparation according to the order product bill of materials,which makes the subsequent material procurement schedule management and procurement schedule control difficult,which will affect the subsequent production,assembly and other links,resulting in project order towing.period.The custom-made equipment for order design and assembly is characterized by a small manufacturing process,most of which are purchased or acquired,and the core components are R&D and assembly.The biggest factor that constrains the progress of such projects is the R&D order of the order and the arrival of the procurement and outsourcing materials required for the assembly.Due to the uncertainties in the capabilities and schedules of all upstream suppliers,the materials required for the project will inevitably lead to delays,which will affect the assembly progress of the orders,resulting in delays in the entire project.In view of the problem that the project arrival delay leads to project delay,this paper will study the material procurement progress forecasting method based on machine learning theory to realize the monitoring of the progress of the required materials of the project.The specific contents are as follows:(1)For the material procurement progress forecasting problem,firstly,according to the enterprise procurement business process,the factors affecting the material procurement progress are analyzed.Then,using the feature engineering method,starting from the original data attributes,combined with the actual situation of enterprise material procurement,extracting and constructing features,and using feature selection methods to reduce the feature subset space.Finally,select a number of different supervised classifiers,use the enterprise real material procurement data,and conduct calculation experiments.The impact of feature engineering on the model prediction results is analyzed and compared,and the material procurement progress forecasting problem is solved.(2)On the basis of the same data and characteristics,use a variety of unsupervised machine learning methods to explore the real data of the enterprise,find the intrinsic characteristics and knowledge in the material procurement progress data,and analyze the material procurement progress forecast from different angles.possibility.(3)Developed a set of material procurement progress and reminder system for customized equipment.By embedding the machine learning prediction model and method studied in this paper into the system,the analysis of the enterprise material procurement data and the prediction of the execution progress can be realized,which can support the procurement personnel's follow-up decision and can be embedded into the enterprise ERP system.This paper focuses on the problem of material procurement progress history data forecasting,using machine learning methods,from data preprocessing,feature engineering to model training and forecasting,to systematically research and apply,to provide decisionmaking for enterprises in the actual procurement progress management process.Certainly intelligent assistance.