Predictive Analytics with AI & Machine Learning PDF

Predictive Analytics with AI & Machine Learning Session 1 What is AI? Any technic which enables computers to Artificial mimic human behaviour Intelligence Learning, vision, speech, etc… Different terms are used: “Machine learning” is now the popularit...

Predictive Analytics with AI & Machine Learning Session 1 What is AI? Any technic which enables computers to Artificial mimic human behaviour Intelligence Learning, vision, speech, etc… Different terms are used: “Machine learning” is now the popularity winner, followed by “artificial intelligence.” “Cognitive computing,” “cognitive technology”. Mimicking certain parts of human intelligence How does AI impact our everyday life?  Social Media Facebook Deep learning is helping Facebook draw value from a larger portion of its unstructured datasets. Most Instagram of its deep learning technology is built on the Torch Instagram uses big data and artificial framework intelligence to target advertising and fight Deep text, translation, Deep Face cyberbullying and delete offensive comments. Chatbots Chatbots recognize words and phrases in order to deliver helpful content to customers who have common questions. Digital Assistants Apple’s Siri, Google Assistant, Amazon’s Alexa, and Microsoft’s Cortana are digital assistants that help users perform various tasks Web Searches Google Predictive Searches: Smart Replies in Gmail When you begin typing a search term Smart replies offer users a way to respond to emails with and Google makes recommendations for simple phrases like “Yes, I’m working on it.” or “No I have not. ” you to choose from, that’s AI in action. with the click of a button. Email Filters in Gmail Google uses AI to ensure that nearly all of the email landing in your inbox is authentic. Generative AI Some examples of AI applications AI in healthcare AI in banking What is AI in reality? It is about Prediction What is AI in reality? Not all automation are AI-based!!! Machine Learning, a subset of AI, that gives Examples: computers the ability to learn without being A model that predicts specifically programmed. fraud in financial Arthur Samuel (1959) transactions. A model that predicts if an Output of machine email is spam or not. learning, is a model which does A model that predicts prediction. (detects) cancer in a radiology scan of a patient.  Training Data Model New Model Output (prediction) Data A B Example Cancer detection Fraud detection A model that Many previous detects new example data frauds (or cancers)  Training Dataset Features Label Model New Data Model Prediction ( e.g. size= 700, bedroom =3) ( e.g. price = 180)  Model A B Model Not cat Some Machine learning terminology Training Data set Testing Dataset Feature AKA independent variable Label/ Target AKA dependent variable  Features Target/ Label Age Education Occupation Gender Income 23 Bachelor’s Engineer Male 50000 > 54 Master’s CEO Female 50000<  Why is More computing power Machine More data Learning so Broad investment: universities, popular governments, startups and tech giants (Google, Amazon, Facebook, these days? Baidu, Microsoft) are all investing heavily in AI.  Understanding Data Types: Cross-Sectional, Time Series, and Panel Data A Brief Overview  Overview of Data Types in Analysis CROSS-SECTIONAL DATA TIME SERIES DATA PANEL DATA DATA COLLECTED AT A DATA COLLECTED AT COMBINES CROSS-SECTIONAL SINGLE POINT IN TIME MULTIPLE POINTS IN TIME, AND TIME SERIES DATA FOCUSING ON ONE ENTITY DATA COLLECTED FROM MULTIPLE ENTITIES ACROSS MULTIPLE PERIODS Understanding Cross-Sectional Data Description: Captures a snapshot of multiple subjects at a single point in time. Useful for comparing differences or analyzing relationships among variables at one time. Example: A survey conducted on 1,000 people in 2024 asking about their internet usage habits. Exploring Time Series Data Description: Example: Involves collecting data points Monthly sales revenue for a for a single subject over time. company from January 2019 to Useful for analyzing trends, December 2023. forecasting, and observing changes over time.  Understanding Panel Data Description: Combines aspects of both cross-sectional and time series data by tracking multiple subjects over multiple time periods. Useful for analyzing changes within subjects over time while also comparing differences between subjects. Example: Annual income data for 100 households collected from 2015 to 2023.