Artificial Intelligence (AI): Any machine which can behave like humans (for a specific task) is a AI Machine.
- There are two ways to achieve it:
- Machine Learning
- Deep Learning
- Types of Machine Learning:
- Supervised ML: works on Labelled dataset for predictions.
- Unsupervised ML: works on unlabelled data. Predictions are not possible, only groups can be created. (clustering)
- Reinforcement
- Supervised ML (Predictions): Prediction can be of two types
- Regression Model (Continuous values)
- Simple LR
- Multiple LR
- Polynomial LR
- Classification Model (Discrete Values)
- Logistic Regression
- K-Nearest Neighbors (kNN)
- Support Vector Machines
- Naive Bayes
- Tree-based Models
- Decision Trees
- Random Forest
- Unsupervised ML
- Clustering
- kMeans
- DBScan
- Association
- Apriori
- Model Evaluation
- Training and Validation
- Model Evaluation Metrics
- Accuracy, RMSE, ROC, AUC, Confusion Matrix, Precision, Recall, F1 Score
- Overfitting and Bias-Variance trade-off
- Regularization (L1/L2)
- Dimension Reduction
- Backward Elimination
- OLS
- Forward Elimination
- Factor Analysis (LDA)
- Principal Component Analysis (PCA)
- Natural Language Processing (NLP)
- Library - NLTK