No abstract

SummaryNowadays, companies want technologies that are able to help them to make the best decision. Data Mining is an excellent tool to estimate the sales. It allows the company to optimize its production and reduce costs, eg, in storage. When these models are combined with Lean Production, it becomes easier to remove waste and optimize industrial production. This case study followed the CRISP‐DM methodology in order to create a model able to reduce and, if possible, eliminate wastage. Several statistics measures were applied to the dataset. Regression algorithms were induced with the goal to find which one of the models are less likely to make mistakes, in other words, what model correctly predict the target result. After executing the tests, the model M1 from the scenario C1 with RandomTree algorithm, average data grouping and average method class creation is the less likely one to give errors regarding regression, having produced an RAE of 6.75%.

Data mining models are an excellent tool to help companies that live from the sales of items they produce because it allows the company to optimize its production and reduce costs, for example in storage. When these models are combined with Lean Production, it becomes easier to remove waste and optimize industrial production. This project is based on the phases of the methodology CRISP-DM, and aims to reduce and, if possible, eliminate wastage. The following methods: average, mean and standard deviation, quartiles and Sturges rule regression, were techniques applied to this data to determine which one is the model is less likely to make mistakes, in other words, meaning that the model did correctly predict the target. Most common metrics used at the statistical level, which had already been proven to have good results in similar studies. After performing the tests, the M4 model is what is less likely to make mistakes in terms of regression with a RAE of 21,33%.