Build Anti-Money Laundering Machine Learning Application

Clarify core terms in data science, computer science, and artificial intelligence. See machine learning as a data science subset, with data mining and big data concepts.

Explore the history of machine learning from Arthur Samuel's checkers program to data-driven models, and how models learn from data, while data cleaning and organization reduce errors.

Explore real-world machine learning applications across speech recognition, medical diagnosis, stock forecasting, product recommendations, spam detection, and fraud prevention, including money laundering detection in digital wallets.

Explore three categories of machine learning: supervised, unsupervised, and reinforcement learning, with examples like car price prediction, K means clustering, and robot navigation with feedback.

Learn money laundering and its three stages: placement, letting, and integration, and how KYC, customer due diligence, and transaction monitoring support anti-money laundering with machine learning to flag suspicious transactions.

Explore pre- and post-event transaction monitoring for anti-money laundering, using customer profiles and rules to flag red flags, generate alerts, and develop typologies into monitoring scenarios.

Explore how machine learning enhances transaction monitoring by detecting evolving money laundering patterns across channels and reducing false positives. Benefit from scalability and cost reductions by identifying highly suspicious transactions.

Explore a hybrid anti-money laundering model based on Simon's four-phase decision making, combining automation and human analysis to monitor client data, detect red flags, and decide on SDR filings.

Develop machine learning algorithms for transaction monitoring using Google Colab and Jupyter Notebook, loading CSV data from Google Drive, and using pandas, numpy, and matplotlib in Python for data exploration.

Explore a Kaggle-sourced, simulated mobile wallet transaction dataset for anti-money laundering analysis, including transaction type, amount, originator, old balance, new balance, destination, and a fraud flag indicating money laundering.

Describe the dataset's types—integers, floats, and objects—and summarize the amount column (min zero, max about 92 million) and show money laundering labels (6 million total, 8213 flagged, 0.12%).

Explore visualization techniques to analyze money laundering data, including missing value checks with heatmaps, box plots, and kernel density estimates, and examine transaction type patterns linked to fraud and thresholds.

Explore correlation analysis with a heat map, normalize and one-hot encode data, and develop a decision tree to detect fraud and money laundering transactions, with train-test split and evaluation.

Address data imbalance by oversampling to balance money laundering and non-money laundering transactions, then train a decision tree and apply a random forest ensemble with majority vote.

Develop and evaluate a random forest classifier on imbalanced and balanced money laundering data, comparing precision and recall while discussing overfitting, underfitting, and dataset limitations.

learn to use tableau desktop for data visualization of bank transactions, using measures and dimensions, filters, and k-means clustering to reveal anti-money laundering patterns.

Explore the concept of ML.NET in a .NET focused machine learning course, and see a movie recommender system demonstration that suggests films based on user preferences.

Explore how recommender systems predict user preferences using content-based and collaborative filtering with movie data. See ML.NET-powered training and a movie recommender app that predicts likelihoods of user choices.

Thank you for purchasing this course; apply the concepts in your real life, and explore the instructor's other machine learning course.