Machine Learning with Business Applications

Programme Overview
Machine Learning algorithms are part of Artificial Intelligence (AI) that imitates the human learning process, which can be used for decision making and problem solving. ML algorithms are systems of problem-solving techniques that exhibit human-like learning capability. While humans learn through practice and experience, machines learn through data. ML algorithms have applications across various industries and different functional areas. The primary objective of ML is to assist in decision making. Today, ML is used for driving innovation and as competitive strategy by several organizations.

The theory of bounded rationality proposed by Nobel Laureate Herbert Simon is evermore significant today with increasing complexity of business problems; limited ability of the human mind to analyze alternative solutions, and the limited time available for decision making. Introduction to Enterprise Resource Planning (ERP) systems has ensured availability of data in many organizations; however, traditional ERP systems lacked data analysis capabilities that can assist the management in decision making. ML assists companies with Robotic Process Automation (RPA) and derives cognitive insights.

Several reports have claimed that AI and Machine Learning specialists in Silicon Valley with few years of experience are paid $300,000 to $500,000 a year1. Bernard Marr, in his article published in the Forbes magazine, claimed that 74% of the customers will be happy to receive computer-generated insurance advice2. While using ML algorithms, we develop several models that can run into several hundreds and each model is treated as a learning opportunity. ML algorithms are classified as follows:
1. Supervised Learning Algorithms,
2. Unsupervised Learning Algorithms,
3. Reinforcement Learning Algorithms, and
4. Evolutionary Learning Algorithms.

In this executive education programme, we discuss various Machine Learning algorithms with their applications using case studies from various industries. The learning pedagogy includes hands-on sessions for better understanding of how ML is used for solving real-life problems.

Programme Objectives:
The course is designed to provide in-depth knowledge of ML algorithms that can be used for fact-based decision-making using case studies from Indian and multinational companies and understand how ML algorithms are used for automation and innovation. Primary objectives of the course are as follows:

• Understand various ML algorithms such as supervised, unsupervised, and reinforcement algorithms.
• Learn to analyse data to gain insights using an appropriate ML algorithm under a given business context.
• Learn various supervised learning algorithms such as regression, logistic regression, decision tree learning, random forest, boosting, neural networks, and deep learning algorithms with applications in solving managerial problems.
• Learn unsupervised learning algorithms such ask-means clustering and factor analysis and its applications.
• Understand how reinforcement and evolutionary algorithms are used by organisations, especially in automation.
• Understand applications of ML in functional areas such as marketing, finance, operations, and supply chain and HR.
• Analyse and solve problems from different industries such as e-commerce, insurance, manufacturing, service, retail, software, banking and finance, sports, pharmaceutical and aerospace using ML algorithms.
• Hands-on experience with software such as Microsoft Excel, Evolver, R, Python, and other proprietary software.

Programme Content:
Supervised Learning Algorithms with Applications in Predictive Analytics:
Simple linear regression: coefficient of determination, significance tests, residual analysis, confidence and prediction intervals. Multiple Linear Regression (MLR): coefficient of multiple coefficient of determination, interpretation of regression coefficients, categorical variables, heteroscedasticity, multicollinearity, outliers, auto-regression and transformation of variables. MLR model development and feature selection. Application of supervised learning in solving business problems such as pricing, customer relationship management, sales and marketing.

Supervised Learning Algorithms with Applications in Classification Problems:
Logistic and Multinomial Regression: Logistic function, estimation of probability using logistic regression, Deviance, Wald test, Hosmer Lemeshow test. Feature selection in logistic regression. Ensemble Methods – Random Forest and Boosting. Business applications of classification problems such as sales conversion, employee attrition, and B2B sales management.

Supervised Learning Algorithms for Forecasting:
Moving average, exponential smoothing, Trend, cyclical and seasonality components, ARIMA (autoregressive integrated moving average), and ARIMAX models. Application of Supervised Learning Algorithms in retail, direct marketing, health care, financial services, insurance, supply chain etc

Unsupervised Learning Algorithms:
Clustering: K-means and Hierarchical

Neural Networks and Deep Learning:
Introduction to Neural Networks: Multilayer perceptron; Backpropagation Algorithms. Deep Learning Algorithms: Convolutional Neural Networks (CNN) and Recrurrent Neural Networks (RNN)

Reinforcement Learning Algorithms:
Markov Chains, Markov Decision Process, Policy Iteration and Value Iteration Algorithms with applications in marketing and finance.

Natural language processing, Text mining and sentiment analysis; Naive Bayes Algorithm.

The following case studies published by the program director at the Harvard Business Publishing will be discussed during the course:

• Package Pricing at Mission Hospitals
• Marketing Head’s Conundrum
• Breaking Barriers – Micro Mortgage Analytics
• Consumer Analytics at Big Basket – Product Recommendations
• Customer Analytics at Flipkart.Com
• Forecasting Demand for Food at Apollo Hospitals
• HR Analytics at Scaleneworks – Behavioural Modelling to Predict Renege
• Predicting Earnings Manipulations by Indian Firms Using Machine Learning Algorithms
• Machine Learning Algorithms to Drive CRM in the online E-commerce site at VMWare
• Consumer choice between house brands and national brands in detergent purchases at Reliance Retail
• Improving lead generation at Eureka Forbes using Machine Learning Algorithms
• Enhancing visitor experience at ISKCON using Text Analytics