Artificial Intelligence - CBSE Class 12 Practical Guide
1. **Key Concepts in AI**:
- Machine Learning (ML)
- Natural Language Processing (NLP)
- Neural Networks
- Data Preprocessing
- Classification, Regression, Clustering
2. **Important Python Libraries**:
- **NumPy**: Used for handling arrays and matrices, supports mathematical functions.
- **Pandas**: Ideal for data manipulation and analysis, provides DataFrame structure.
- **Matplotlib/Seaborn**: Libraries for data visualization (graphs, charts, plots).
- **Scikit-learn**: Essential for implementing machine learning algorithms like regression,
classification, clustering.
- **TensorFlow/PyTorch**: Advanced libraries for deep learning, neural networks, and complex
models.
- **Keras**: A high-level neural networks API, works on top of TensorFlow.
3. **Steps in AI Model Building**:
1. Data Collection
2. Data Preprocessing:
- Handle missing values, normalization, encoding categorical variables.
3. Splitting Data: Train and test datasets.
4. Model Selection: Choose the right algorithm (e.g., Linear Regression, Decision Trees).
5. Training the Model: Train using training data.
6. Evaluation: Assess model performance using metrics (Accuracy, Precision, Recall, etc.)
7. Hyperparameter Tuning: Optimize model parameters.
4. **Common Algorithms to Practice**:
- **Linear Regression**: For predicting continuous values.
- **Logistic Regression**: For binary classification tasks.
- **Decision Trees**: A flowchart-like structure for decision-making.
- **K-Nearest Neighbors (KNN)**: A simple classification algorithm based on nearest data points.
- **Support Vector Machine (SVM)**: For classification tasks in higher dimensions.
- **Naive Bayes**: A probabilistic classifier based on Bayes' theorem.
- **Random Forest**: An ensemble method combining multiple decision trees.
5. **Evaluation Metrics**:
- **Accuracy**: The percentage of correct predictions.
- **Confusion Matrix**: A matrix that shows the performance of a classification model.
- **Precision, Recall, F1-Score**: Metrics for evaluating classification models, especially
imbalanced datasets.
- **Cross-Validation**: Used for assessing how the model generalizes to unseen data.
6. **Recommended Tools**:
- **Jupyter Notebook**: Popular tool for running Python code interactively and visualizing results.
- **Google Colab**: Cloud-based platform for running Python code, great for heavy tasks.
- **IDE**: PyCharm, VS Code for efficient coding and debugging.
7. **Practical File Organization**:
- Keep your practical file well-organized with clear headings and steps for each experiment.
- Include the problem statement, code, output, and a brief explanation of the approach used.
- Document all models and algorithms you have implemented.
8. **Project Ideas**:
- **Predict House Prices**: Implement a linear regression model.
- **Image Classification**: Use neural networks to classify images.
- **Sentiment Analysis**: Use NLP techniques for analyzing text sentiments.
- **Recommendation System**: Build a simple recommendation system.
9. **Viva Voce Preparation**:
- Be clear about the algorithms you have implemented, why you chose them, and how they work.
- Be ready to explain concepts like overfitting, underfitting, bias-variance tradeoff.
- Practice explaining the steps involved in building models and evaluating them.
