Week 6: Data Cleaning

• Missing value detection and imputation techniques
• Error handling in Python (try-except, exceptions)
• File handling and data I/O operations
• Code documentation (docstrings) and modularization
• Building reusable data cleaning functions

Week 7: Outlier Treatment & Feature Engineering

• Outlier detection (Z-score, IQR, Isolation Forest)
• Feature engineering from datetime and RFM analysis
• Aggregation and time-based features
• Mathematical transformations (log, sqrt, polynomial)
• Creating interaction features

Week 8: Feature Encoding & Scaling

• Categorical encoding (Label, One-Hot, Target, Frequency encoding)
• Feature scaling (Standardization, Min-Max, Robust scaling)
• Column standardization strategies
• Creating preprocessed datasets
• Feature transformation pipeline documentation

Week 9: Feature Selection & Math Foundations

• Feature importance concepts and curse of dimensionality
• Filter methods (correlation, variance threshold, chi-square)
• Wrapper and embedded methods for selection
• Matrix algebra: vectors, matrices, eigen values/vectors
• Train-test-validation split with stratified sampling

Week 10: ML Fundamentals & Metrics

• Types of ML (Supervised, Unsupervised, Reinforcement)
• Classification vs Regression and validation techniques (K-fold, Stratified CV)
• Evaluation metrics (Accuracy, Precision, Recall, F1, ROC-AUC)
• Imbalanced dataset handling (SMOTE, class weights)
• Bias-variance tradeoff

Week 11: Probability & Linear Models

• Probability basics, conditional probability, Bayes' rule
• Probability distributions (Normal, Bernoulli, Binomial)
• Linear regression and Logistic regression
• Regularization (L1, L2, Elastic Net)
• Model interpretation and coefficient analysis

Week 12: Distance & Probabilistic Models

• K-Nearest Neighbors (KNN) algorithm and distance metrics
• Optimal K selection and feature scaling impact
• Naïve Bayes classifier (Gaussian, Multinomial, Bernoulli)
• Model comparison and performance analysis
• Selecting best performing model

Week 13: Tree-Based Models & SVMs

• Decision Trees (CART, entropy, Gini impurity, pruning)
• Support Vector Machines (linear, kernel trick, RBF)
• Hyperparameter tuning for trees and SVMs
• Decision boundary visualization

Comprehensive traditional ML model comparison

Week 14: Bagging & Random Forest

• Ensemble learning principles and bagging technique
• Random Forest algorithm and feature importance
• Hyperparameter tuning (n_estimators, max_depth, max_features)
• Out-of-bag error estimation
• Voting classifiers (hard and soft voting)

Week 15: Boosting Methods

• Boosting fundamentals (AdaBoost, Gradient Boosting)
• XGBoost, LightGBM, CatBoost overview
• Hyperparameter optimization (learning_rate, max_depth, regularization)
• Stacking and blending ensembles
• Ensemble method comparison

Week 16: Statistical Hypothesis Testing

• Population vs sample, Central Limit Theorem
• Confidence intervals and standard error
• Hypothesis testing (null/alternative, p-values, significance)
• Z-test and t-test for model comparison
• Cross-validation strategies and model selection

Week 17: Model Evaluation & Deployment Prep

• Converting ML metrics to business KPIs
• ROI calculation and cost-benefit analysis
• Model monitoring and A/B testing design
• Model documentation and deployment artifacts
• Final churn prediction model with business impact report

Week 18: Neural Networks Foundation

New Project: Customer Review Sentiment Analysis

• Biological vs artificial neurons, Perceptron and limitations
• Multilayer Perceptron (MLP) architecture
• Activation functions (Sigmoid, Tanh, ReLU, Softmax)
• Forward propagation, loss functions, backpropagation
• TensorFlow/Keras for building neural networks

Week 19: Deep Learning Optimization

• Gradient descent variants and optimizers (SGD, Adam, RMSprop)
• Learning rate scheduling and batch size selection
• Regularization (Dropout, Batch Normalization, Early Stopping)
• Vanishing/exploding gradient problems
• Hyperparameter tuning with TensorBoard monitoring

Week 20: NLP Fundamentals

• Text preprocessing (tokenization, stemming, lemmatization, stop words)
• Bag of Words and N-grams (unigram, bigram, trigram)
• Count Vectorizer and TF-IDF vectorization
• Text classification using TF-IDF + ML models
• Sentiment analysis baseline model

Week 21: Word Embeddings & RNNs

• Word2Vec (CBOW and Skip-gram) and pre-trained embeddings
• Recurrent Neural Networks (RNN) and LSTM/GRU architectures
• Backpropagation through time (BPTT)
• Bidirectional RNNs for sequence modeling
• LSTM-based sentiment analysis model

Week 22: Large Language Models (LLMs)

New Component: LLM-powered Chatbot

• Transformer architecture and self-attention mechanism
• Pre-trained LLMs (BERT, GPT) and Hugging Face Transformers
• Fine-tuning BERT for sentiment classification
• Prompt engineering and few-shot learning
• Building chatbot with LLMs and RAG (Retrieval Augmented Generation)

Week 23: Time Series Fundamentals

New Project: Retail Sales Forecasting

• Time series components (trend, seasonality, cyclical, residual)
• Stationarity testing (ADF test) and differencing
• Autocorrelation (ACF) and Partial Autocorrelation (PACF)
• Time series decomposition methods
• Sales data exploration and pattern identification

Week 24: Time Series Forecasting

• ARIMA models (AR, MA, ARMA, ARIMA, SARIMA)
• Model selection using AIC/BIC and statsmodels library
• LSTM for time series forecasting
• Multi-step ahead forecasting
• Comparing statistical vs deep learning approaches

Week 25: Unsupervised Learning

New Project: Customer Segmentation

• Clustering algorithms (K-Means, DBSCAN, Hierarchical)
• Clustering evaluation metrics (Silhouette, Davies-Bouldin)
• Principal Component Analysis (PCA) and covariance matrix
• Singular Value Decomposition (SVD) and t-SNE
• Customer segmentation with profiling and insights

Week 26: Capstone Project - Phase 1

• Capstone project selection and business problem definition
• Data collection from multiple sources
• Comprehensive EDA and data quality assessment
• Data cleaning, feature engineering, and feature selection
• Train-test split preparation

Week 27: Capstone Project - Phase 2

• Building baseline and advanced ML models
• Hyperparameter tuning and model optimization
• Cross-validation and statistical model selection
• Business impact analysis and ROI calculation
• Model documentation and deployment strategy

Week 28: Final Presentations & Portfolio

• Complete project documentation and GitHub repository
• Portfolio website/showcase development
• Presentation preparation with storytelling
• Final project presentations and Q&A
• Career guidance (resume, LinkedIn, interview prep)