Diseño de un modelo predictivo, mediante la técnica de minería de datos para la asignación de recursos en la producción de café solido soluble para calidad A/R de la compañía ASKELGADO S.A.

Abstract

The present qualification work of the development of a design of a predictive model for the allocation of resources in the production of soluble solid coffee for A / R quality of the company ASKELGADO S.A. aims to identify the resources to be used, recognize a data mining technique, use a predictive model, and evaluate the model for A / R quality. The research used was of the documentary type, since an interview was conducted with the 2 people in charge of the production to find out their current status regarding the allocation of resources. For the project, the design of a predictive model was proposed, using the data mining technique to identify resources through historical data. To carry out the research, a qualitative, analytical research approach was used to study the context where the problem of resource identification exists; The decision trees technique are supervised learning techniques, in which functions are learned, relationships that associate inputs with outputs, so they conform to a set of examples of which we know the relationship between the input and the desired output. The mining technique is supervised, for that python will be used and the methodology is CRISP-MD. A tree was designed that generated a predictive evaluation model for decision making in the plant. In the end, the recommendations were made to be considered as the improvement of the previously used model.