Prediction Problems Of Mould Delivery Date Research Based On Data Mining

Model owns a very important place in contemporary industry. Whether the enterprises can deliver the products to the customers on time, it is a important part of enterprise production management work and one of enterprise’s core competitive force. However, the delivery delay happens very often due to various relative factors during the model process. Estimating a relative accurate mold delivery date, it is very important to cut down tardiness rate, at present, it is mainly through experience to forecast mold delivery date and short of a scientific approach.The thesis basing on BP artificial neural networks proposes a predicting way of unfinished mold orders using historical data of enterprise.Firstly, through in-depth company research, summarized and analyzed the factors that may affect the delivery date of the mold, it can be divided two parts:dynamic factor and static factor. For static factor, it mainly includes Number Sets Of Mold Per Order, Product Type and Rated Labor-hour Per Order etc. For dynamic factor, it contains the progress and the load etc.Secondly, to extract some data about production from the background enterprise database, in order to selecting the real significative factors, use a data analysis way to do Correlation Analysis for these factors. Putting the selected effective factors as input variables and the remaining completion time of a mold order as output variables to establish the BP neural network model.The following, by altering model training algorithm, the number of hidden layers and hidden layer nodes of the network model to conduct lots of training. After enough of training, to verify the reliability of the model using test data sets.Finally, applying the software engineering’s method to develop and design a software system of predicting mold’s order delivery date. This system have been in use among the actual enterprise production and can help Business people and Production planners Fast forecast Mold-order’s delivery date.In summary, The thesis use BP neural network theory, building a unfinished mold orders prediction model. Develop and design a software system of predicting mold’s order delivery date.