Sistema de soporte de decisiones para la producción porcina, aplicando técnicas de minería de datos e inteligencia de negocios

Abstract

The importance of using a good decision indicates that a problem or situation is valued and considered to better choose the path to follow according to the different alternatives or options. Currently, pig production in the country has been increasing; for its continued growth, making a good decision will be crucial for its persistence in the market. Support for decision-making in pig production will depend on the proper management of data and information generated in the company, and in a certain way, data mining (DM) and business intelligence (BI) techniques are significant support within the company; allowing the entrepreneur to know and understand his own business and apply improvements and make decisions that mean a competitive advantage. In the administration of the data there are certain inconsistencies on the part of the company under study located in the canton of Balsas, province of El Oro, such as: non-standard formats, disorganization of the information, disorder and failures in the manual processing and interpretation of data. the data, causing effects on pig production; in addition to having an unwise decision making in the face of the current demand. The objective of this work is to implement a decision support system (SSD) in the company PORCIPRO (fictitious name for data privacy reasons), through business intelligence and data mining, which contributes to the descriptive and predictive analysis of pig production, to the owner of the business, this will allow the company to continue in the market and even increase its production. , The company's data has been selected and extracted from different sources, to later be transformed and organized into files with the appropriate format; in such a way that the SSD processes and analyzes the data of said files, allowing the visualization of the resulting information in different statistical graphs presented in a web interface with a BI dashboard style. In the system, the user can select different modules; In module 1 in relation to the number of live pups per mother, years, and period; It also allows predicting trends in offspring production, through the application of predictive data mining techniques, such as time series. In module 2 of expenses, all production costs made by the company per year are presented; In module 3 of sales, the income is presented by years. And in the last module 3 the profits of the company by year are shown. Each module has its information boxes at the top as a detail of the graph. In the development of this work, the DataOps methodology was applied, which more than a methodology, is a work style that the company must opt for to improve data processing. The results obtained in the work are based on the timely detection of the state of the company and future projection based on the key performance indicators of pig production; In addition, improvements in the business and significant changes can be observed in terms of the organization and the actions that are carried out in the pig production process, supported by the statistical information provided by the SSD that was developed in this work.