AI & Machine Learning – Complete Self-Learning Syllabus

This is an introduction to AI & Machine Learning self-learning. If you are a newcomer or do not have a strong foundation in basic programming, please complete basic programming fundamentals C, Java & OOPS first before starting ML.

---

0. Orientation & Foundations

---

0.1 Artificial Intelligence (AI)

Overview

• What is Artificial Intelligence

• Ability of machines to mimic human intelligence

• Real-world AI examples

---

0.2 Machine Learning (ML)

Overview

• What is Machine Learning

• Subset of Artificial Intelligence

• Uses data to solve tasks

• Learns patterns from past data

---

0.3 Deep Learning (DL)

Overview

• What is Deep Learning

• Subset of Machine Learning

• Uses neural networks inspired by the human brain

---

0.4 Comparison of Concepts

AI vs ML vs DL

• AI as a broad concept of intelligent systems

• ML as data-driven statistical learning

• DL as neural-network-based learning

ML vs Data Roles

• Machine Learning vs Data Science

• Machine Learning vs Data Analyst

---

0.5 Traditional Programming vs Machine Learning

Traditional Programming

• Rules + Data → Output

• Fixed logic

• No learning from data

Machine Learning

• Data + Output → Model

• Model learns rules automatically

• Improves with experience

---

0.6 Usage of Machine Learning

Why Machine Learning is Used

• Handles large amounts of data

• Works with structured and unstructured data

• Learns automatically without explicit rules

• Improves performance over time

When Machine Learning Should NOT Be Used

• Very small datasets

• Simple rule-based problems

• No clear objective

Real-World ML Systems Overview

• Recommendation systems

• Fraud detection systems

• Quality control systems

• Autonomous decision systems

---

1. Introduction to Machine Learning

---

1.1 Basics

Definition & Concept

• Definition of Machine Learning

• How machines learn from data

• Learning from historical data

• Predictive capability

• Examples of ML in daily life

ML Models

• Trained using data

• Based on probability, statistics, and linear algebra

---

1.2 Data in Machine Learning

Why ML handles Data

• Handles large amounts of data

• Improves performance with experience

Real-World Data Examples

• Structured data (Rows, columns, databases)

• Unstructured data (Text, images, audio)

• E-commerce data (Sales reports)

• Customer datasets (Age, Gender, Location)

---

2. Types of Machine Learning & Algorithms

---

2.1 Supervised Learning

Overview

• Definition: Uses labeled data where input and output are known

• Target variable is known

2.1.1 Regression

• Definition: Predicts continuous values

• Examples: House price prediction, Temperature prediction, Stock prices

• Algorithms:

  • Linear Regression

  • Multiple Linear Regression

• Evaluation Metrics:

  • Mean Squared Error (MSE)

  • R² Score

2.1.2 Classification

• Definition: Predicts categorical values

• Examples: Spam detection, Disease diagnosis, Pass/Fail

• Algorithms:

  • Logistic Regression

  • K-Nearest Neighbors (KNN)

  • Decision Tree

  • Random Forest

  • Support Vector Machine (SVM)

  • Naive Bayes

• Evaluation Metrics:

  • Accuracy

  • Precision

  • Recall

  • F1-Score

  • Confusion Matrix

  • Classification Report

---

2.2 Unsupervised Learning

Overview

• Definition: Uses unlabeled data to find hidden patterns

2.2.1 Clustering

• Definition: Groups similar data points

• Examples: Customer segmentation, Product grouping

• Algorithms:

  • K-Means Clustering

  • Hierarchical Clustering

2.2.2 Dimensionality Reduction

• Definition: Reduces number of features for visualization and performance

• Algorithms:

  • Principal Component Analysis (PCA)

---

2.3 Semi-Supervised Learning

Overview

• Uses combination of labeled and unlabeled data

• Used when labeling is expensive

---

2.4 Reinforcement Learning

Overview

• Learns using trial and error

Components

• Agent

• Environment

• Actions

• Rewards

• Policy

Applications

• Game playing

• Robotics

• Autonomous systems

---

2.5 Comparison of ML Types

Analysis

• Differences between supervised, unsupervised, semi-supervised, and reinforcement learning

• Use-cases for each ML type

---

3. Applications of Machine Learning

---

3.1 Industry Use-Cases

Key Areas

• Recommendation systems

• Image recognition

• Speech recognition

• Natural Language Processing (NLP)

• Fraud detection

• Healthcare

• Manufacturing & quality control

• Autonomous systems

• Chatbots & virtual assistants

---

4. Machine Learning Workflow

---

4.1 End-to-End ML Pipeline

Steps

1. Data Collection

2. Data Pre-processing

   • Cleaning the data after collecting it

     • Handling missing values

     • Removing duplicate values

     • Handling other anomalies such as skewed data, outliers, noise, etc.

3. Exploratory Data Analysis (EDA)

   • Understanding and studying the data

   • Gaining strong knowledge about the dataset

   • Analyzing data distributions, relationships, and patterns

4. Feature Engineering

   • Creating or adding new columns (features) into the dataset if required

   • Feature Encoding:

     • In the data, there might be categorical values (for example, string data) These need to be converted into numerical values.

     • To do this conversion, many encoding methods are available in machine learning

       • One-Hot Encoding

       • Dummy Encoding

       • Label Encoding

       • etc.

5. Feature Selection

   • The dataset may contain many unnecessary or unwanted columns

     • Feature selection is the process of selecting only the necessary columns

     • Many machine learning algorithms are available to perform feature selection

6. Split into Training and Testing Sets

   • 80% of the data is used for training

   • The remaining 20% of the data is used for testing

   • The same data should not be used for both training and testing

7. Feature Scaling

   • The dataset may contain values in different units or formats

     • Making uniformity among these values is called feature scaling

   • Techniques used for feature scaling in machine learning include

     • Standard Scaling

     • Min-Max Scaling

     • etc.

8. Building the Machine Learning Model

   • In machine learning, there are many algorithms such as regression and classification

     • Linear Regression

     • Logistic Regression

     • Clustering

     • etc.

   • After understanding the problem and task, an appropriate ML algorithm is selected to build the model

9. Model Evaluation

   • After building the model, it must be tested or evaluated

     • Various model evaluation metrics are used to measure performance

10. Hyperparameter Tuning

    • If the model performance is not sufficient or the result is not good, it is improved by providing more training or adjust parameters to improve performance

11. Model Saving

    • Once the model performance is verified as good during evaluation

      • The trained model is saved (Machine learning libraries provide methods to save models).

12. Testing with Unseen Data

    • After completing all processes, the model is tested with unseen data

    • Training and evaluation are done using available data. Fresh, new data used for testing is called unseen data

13. Model Deployment

    • This is the final stage of the machine learning workflow

    • The trained model is implemented in a real-world application

---

4.2 Data Preprocessing

Tasks

• Handling missing values

• Handling outliers

• Removing duplicates

• Encoding categorical variables

• Feature scaling

  • Normalization

  • Standardization

---

4.3 Exploratory Data Analysis (EDA)

Techniques

• Dataset overview

• Data types and shape

• Statistical summary

• Data distribution

• Central tendency

• Data spread

• Correlation analysis

• Visualization

---

5. Python Programming for Machine Learning

---

5.1 Python Basics

Fundamentals

• Python introduction

• Variables

• Keywords

• Comments

• Indentation

---

5.2 Data Types

Types

• Integer

• Float

• String

• Boolean

• Type conversion

---

5.3 Data Structures

Structures

• List

• Tuple

• Set

• Dictionary

• Differences between data structures

• Use-cases of data structures in ML

---

5.4 Operations on Data Structures

Operations

• Insert

• Update

• Delete

• Indexing

• Slicing

---

5.5 Operators & Control Flow

Operators

• Arithmetic operators

• Relational operators

• Logical operators

• Assignment operators

• Membership operators

Control Statements

• if

• if-else

• elif

• for loop

• while loop

• break

• continue

• pass

• Difference between for loop and while loop

---

5.6 Functions & IO

Inbuilt Functions

• len()

• sum()

• min()

• max()

• sorted()

• type()

Input & Output

• input()

• print()

• Formatted output

---

5.7 Logical Practice Problems

Practice

• Prime number

• Palindrome

• Armstrong number

• Fibonacci series

• Factorial

• Unique elements in list

• Frequency counting

• Largest and smallest element

• Pattern problems

• Array and string problems

---

6. Python Libraries for Machine Learning

---

6.1 NumPy

Concepts

• Arrays

• Array operations

• Vectorized operations

• Mathematical functions

---

6.2 Pandas

Concepts

• Series

• DataFrame

• Data loading

• Data cleaning

• Data manipulation

---

6.3 Data Visualization

Concepts

• Matplotlib

• Seaborn

Plots

• Line plot

• Bar plot

• Histogram

• Box plot

---

6.4 Scikit-Learn

Concepts

• Introduction to scikit-learn

• Datasets

• Model training

• Model prediction

• Model evaluation

---

7. Statistics for Machine Learning

---

7.1 Descriptive Statistics

Measures

• Mean

• Median

• Mode

• Range

• Variance

• Standard deviation

• Quartiles

• Interquartile range (IQR)

---

7.2 Statistical Equations

Formulas

• Mean formula

• Variance formula

• Standard deviation formula

• Quartile calculation

---

7.3 Usage of Statistics in ML

Applications

• Mean for normalization

• Median for outlier handling

• Variance for feature importance

• Standard deviation for scaling

• Quartiles for data distribution analysis

---

8. Mathematics for Machine Learning

---

8.1 Vectors

Concepts

• Vector definition

• Vector representation

• Vector addition

• Scalar multiplication

• Dot product

• Vector usage in ML

---

8.2 Matrices

Concepts

• Matrix representation

• Matrix addition

• Matrix multiplication

• Matrix transpose

• Identity matrix

• Matrix usage in ML

---

9. Probability for Machine Learning

---

9.1 Probability Basics

Concepts

• Probability definition

• Sample space

• Events

---

9.2 Types of Events

Categories

• Independent events

• Dependent events

• Conditional probability

---

9.3 Probability in Machine Learning

Applications

• Classification problems

• Prediction confidence

• Risk and uncertainty

• Naive Bayes intuition

---

9.4 Advanced Probability Topics

Topics

• Bayes theorem

• Random variables

• Probability distributions

• Normal distribution

• Binomial distribution

---

10. Model Evaluation & Optimization

---

10.1 Evaluation Concepts

Topics

• Training data vs testing data

• Overfitting

• Underfitting

• Bias vs variance

• Cross-validation

• Hyperparameter tuning

---

11. Machine Learning Projects

---

11.1 Project List

Projects

• Student performance prediction

• House price prediction

• Customer segmentation

• Spam detection

• Recommendation system (basic)

• Quality defect prediction

• End-to-end ML project workflow

---

12. MLOps & Deployment (Introductory)

---

12.1 Deployment Basics

Concepts

• Model saving and loading

• Basic deployment concepts

• ML lifecycle overview

• Monitoring models

---

13. Ethics & Responsible AI

---

13.1 Responsible AI

Topics

• Bias in ML models

• Fairness

• Explainability

• Privacy concerns