Course Outline

Module 1: Introduction to Machine learning

• Applications of Machine Learning
• Making accurate prediction and working with robust Machine learning models
• Working with categorical and missing data
• Splitting data into Training set and testing set
• Feature scaling and Data preprocessing

Module 2: Introduction to Fraud Detection

• The basics of fraud
• Types of frauds against Banking
• Understanding fraud Detection, the traditional method and the algorithms used
• Introduction to packages and python libraries to use such as Pandas, Scikit-Learn and Numpy
• Understand the SMOTE library
• Supervised algorithm in ML to catch a fraud
• Un-Supervised algorithm in ML to catch a fraud
• Work with logistic Regression
• Use of pipeline

Module 3: Using KNN for Fraud Detection

• Working with unlabeled data
• Comparison of Normal and Abnormal Behavior
• Data exploration and customer segmentation
• Implementation of Naives Bayes
• Working with the Elbow method
• Assigning fraud versus non-fraud
• Detection of Outliers and Residuals

Model 4: Using Random forest for fraud detection

• Working with Unlabeled data
• Understand Random forests classifiers
• Performing evaluation and metrics on the model
• Plotting the recall Curve and adjusting the Algorithm weights
• Adjust the Random Forest for fraud detection
• Understanding the Ensemble method
• Using logistic regression and Voting classifiers

Module 5: Unsupervised Learning for Fraud Detection

• Understand the K-Mean clustering
• Work with Hierarchical clustering
• Understand Density-based Clustering