Week 2: Python for ML

Python Basics

• Introduction to Jupiter Note book / Google collab
• Data Types (int, float, str, bool)
• Creating Variables
• Arithmetic, comparison, Logical Operators

Control Structures & Functions

• if-else conditions
• if-elif-else condition
• Nested if condition
• for loops
• range() , Functions (pre-defined, user defined) expressions

Python Libraries

• NumPy arrays and vectorized operations
• Matplotlib basic plotting
• Seaborn for visual enhancement
• Real-world EDA practice

Data Cleaning with Pandas

• Reading and inspecting data
• Handling missing values (dropna, fillna)
• Removing duplicates
• Data type conversion
• Filtering, grouping, merging datasets

Week 3: Machine Learning – Part 1

ML Pipeline Overview

• ML lifecycle and workflow
• Supervised vs. Unsupervised learning
• Train/Test split using sklearn
• Overview of sklearn pipeline creation
• Practical setup for ML project

Regression Algorithms

• What is meant by Model
• Linear Regression (Simple & Multiple)
• Regression evaluation metrics (MSE, RMSE, R^2)
• Hands-on: Predicting house prices

Classification Algorithms

• Logistic Regression fundamentals
• K-Nearest Neighbors (KNN)
• Classification use cases (e.g., cancer detection)
• Hands-on with sklearn models

Model Evaluation Techniques

• Confusion Matrix
• Precision, Recall, and F1-Score
• Cross-validation techniques
• GridSearchCV basics

Week 4: Machine Learning – Part 2

Tree-Based Models

• Decision Tree algorithm
• Splitting Criteria: Gini vs Entropy
• Random Forest ensemble learning
• Feature importance ranking
• Overfitting and pruning techniques

SVM & XGBoost

• Support Vector Machines: linear and non-linear
• Kernel trick explained
• Introduction to Gradient Boosting
• XGBoost installation and model tuning
• Comparison: RF vs SVM vs XGBoost

Unsupervised Learning

• K-Means Clustering: Elbow Method
• Hierarchical Clustering
• Distance metrics: Euclidean
• Applications of clustering in real-world problems
• Hands-on clustering mini project

Feature Engineering

• One-hot and Label Encoding
• Feature Scaling: Standard vs MinMax
• Feature selection techniques

Week 5: Natural Language Processing – Part 1

Introduction to NLP

• NLP pipeline overview
• Applications: chatbots, sentiment analysis, etc.
• Challenges of NLP
• Text encoding techniques overview
• Introduction to nltk

Text Cleaning & Tokenization

• Removing punctuation, numbers, stop words
• Lemmatization vs Stemming
• Tokenization using nltk
• POS tagging basics
• Case study: cleaning feedbacks data

Text Vectorization

• Bag of Words (BoW)
• N-grams and context windows
• TF-IDF implementation
• Sparse matrix representation
• Visualization with word clouds

Text Classification

• End-to-end NLP pipeline using Scikit-learn
• Dataset splitting and vectorization
• Training Logistic Regression or Naive Bayes
• Real-time sentiment prediction demo

Week 6: Natural Language Processing – Part 2

Word Embeddings

• Word2Vec: Skip-gram & CBOW
• GloVe embeddings
• Cosine similarity between words
• Visualizing word vectors
• Hands-on: Pre-trained embeddings usage

Sequence Models and RNN Basics

• Sequential data and temporal relationships
• Introduction to RNNs and limitations
• Tokenizing and padding sequences
• Preparing text data for DL
• Keras Embedding layer overview

Transformers & Hugging Face

• Introduction to transformers
• Pretrained models: BERT, GPT
• Hugging Face installation & usage
• Text classification pipeline with Transformers
• Fine-tuning basics

NLP Capstone Project

• Select a domain-specific problem (e.g., reviews, healthcare)
• Dataset analysis and cleaning
• Model selection and evaluation
• Presenting the outcome
• Review and feedback

Week 7: Deep Learning – Part 1

Neural Networks Fundamentals

• Perceptron model and activation functions
• Feedforward NN architecture
• Forward and backward propagation
• Loss functions: MSE, Cross-Entropy
• Model intuition with simple examples

Training Deep Models

• Epochs, Batches, and Batch Size
• Optimizers: SGD, Adam
• Learning Rate tuning
• Underfitting vs Overfitting
• Hands-on: Build a basic NN using Keras

Keras/TensorFlow Practice

• TensorFlow/Keras basics
• Model.compile(), .fit(), .evaluate()
• Evaluate performance on test data

Hyperparameter Tuning

• Manual tuning vs Grid Search
• Dropout regularization
• Learning rate schedulers
• Case Study: Improve model performance

Week 8: Deep Learning – Part 2

CNNs for Image Classification

• Convolution and pooling operations
• CNN architecture and feature extraction
• Activation maps
• Building CNNs with Keras
• MNIST image classification project

Transfer Learning

• Concept of Transfer Learning
• Pre-trained models: VGG16, ResNet50
• Feature extraction vs Fine-tuning
• Load, modify and evaluate pre-trained models
• Image classification case study
  DL with NLP (Transformers)
• Transformer architecture recap
• BERT for sequence classification
• Tokenizer and model pipeline
• Evaluate model performance
• NLP fine-tuning best practices

Deep Learning Capstone Project

• Choose between NLP or Image-based project
• Define objective and strategy
• Model training and validation
• Wrap-up and guidance for next steps