Real-time business intelligence development using machine learning to increase the potential of the dairy goat milk business

Abstract

The development of big data and real-time data warehouse (RTDW) technologies has transformed traditional business intelligence (BI) into real-time business intelligence (RTBI). The RTBI framework is developed in this study by incorporating machine learning-based real-time prediction features. The complexity of layer integration in the RTBI framework is a challenge in building RTBI. The development of RTBI was carried out in business areas that did not have RTBI from the beginning, such as the dairy goat milk business in Probolinggo, East Java. Another main reason is that the dairy goat milk business is a food alternative to cow's milk in Indonesia. The results of this study can contribute to increasing the potential value of the goat milk business. The research method was developed by adapting to the Kimball method and unified modeling language (UML). The real-time prediction feature with the long short-term memory (LSTM) algorithm is the main feature in the RTBI framework developed in the research. The calculation results of real-time predictive analysis latency successfully approached 0 milliseconds (ms), namely 9.35×10^-5 ms. The application of RTBI in the dairy goat milk business was successfully built but the real data is very limited, so RTBI is less able to describe the movement of the business.