Performance measures forecasting applying Machine Learning methods

Abstract

The goal of this thesis is to improve performance management and business choices applying Machine Learning (ML) methods, exploiting the opportunities offered by Big Data. In fact, Big Data with the support of Business Intelligence (BI) and Business Analytics (BA) may be utilized to analyze large volumes of data in order to build machine learning algorithms to assist organizations to make better and faster decisions. In particular, building a forecasting model for a KPI or for other performance measures is useful because helps to plan and manage the business, to discover the influence exercisable and find leverage to improve the performance measurements. Moreover, the purpose is to analyze the outgoing orders using Machine Learning to calculate and predict a date for the actual delivery, allowing to optimize the choices related to orders and, consequently, overall business performance. Therefore, having insight into future KPI or other performance measures trends provides opportunities for business managers to take preventative actions that can circumvent unwanted business situations. In this thesis, the importance of Performance Management and Business Intelligence (BI) is analyzed and how machine learning methods can help the areas mentioned. The aim is the study of an AutoML model and an additional framework for analyzing and forecasting performance measures by the application of more recent machine learning and deep learning techniques.