DATA SCIENCE AI & ML WITH GEN AI

Machine Learning

Data Preprocessing Techniques

• Data Imputation (Missing values)
• Simple Imputation
• KNN Imputation
• Iterative Imputation
• Data Encoding Techniques
• Label Encoding
• OneHot Encoding
• Finding Outliers
• IQR technique
• Zscore technique
• Percentile technique
• LocalOutlierFactor
• Data Normalization, Transformation, Scaling
• MinMaxScaler
• StandardScaler
• RobustScaler
• PowerTransform
• Box-CoxTransform
• QuantileTransforms
• Dimensionality Reduction Techniques
• PCA- Principle Component Analysis
• SVD- Singular Value Decomposition
• LDA- Linear Discriminant analysis
• Feature Selection (Importance) and Engendering techniques
• Supervised learning based
• Unsupervised learning based
• feature importance interpretation (SHAP, LIME)
• Case Studies on Data Preprocessing techniques and comparative analysis of various techniques

Regression Analysis

• Regression, Linear regression
• Linear regression, Multiple Regression
• Ridge Regression, Lasso Regression
• Explanation of statistics
• Evaluation metrics (R-Squre, Adj R-Squre, MSE, RMSE)
• Train/Test Split, Hypothesis testing formal way
• Case Studies
• Project on Regression Analysis from Kaggle

Classification

• Introduction to Machine Learning
• Naïve Bayes classifier
• Decision Tree classifier
• KNN classifier
• Logistic Regression
• Support Vector Machines (SVM)
• One-vs-Rest and One-vs-One for Multi-Class Classification
• Predict() Vs PredictProba()
• Ensemble models (Random Forest, Bagging, Boosting)
• Bagging algorithms
• Boosting algorithms
• Stacking algorithms
• Xgboost indepth with Industry cases
• SK Learn ML library using Python and Case Studies
• Project on Classification Algorithms from Kaggle

Model Selection and Evaluation

• Accuracy measurements
• Precision, Recall, Precision – Recall Trade-off
• AUC Score, ROC Curve
• Train/Validation/Test split, K-Fold Cross Validation
• The Problem of Over-fitting (Bias-Variance Trade-off)

Learning Best Practices for Model Evaluation

• Bias, Variance, Overfitting, Underfitting methods
• Pipelining
• Parameter Tuning mechanisms (Grid Search, Random Search)
• Debugging algorithms with learning and validation curves

Case Study

Association Analysis

• Association Rules & Interesting measures
• Apriori Algorithm
• FP-Growth algorithm
• Case Studiies

Clustering

• Similarity distance measures
• K-means Clustering
• Hierarchical Clustering
• DB Scan Clustering
• Case Studies
• MLOPS
• MLFLOW , End to End ML case Studies
• DVC

Applications of ML

• Time Series Analysis (Stock Market forecasting using ARIMA models)
• Recommendation Systems (Filter based RS and Collaborative based RS)
• Dealing with Imbalanced datasets (Anomaly Detection Methods)
• Deployment ML/DL/NLP models using Flask, Fast API, Github
• ML Model Deployment in Azure