ML-Notebook

Karma of Humans is A.I.

A comprehensive, hands-on machine learning curriculum — from NumPy basics to neural networks. Each module contains Jupyter notebooks with from-scratch implementations alongside scikit-learn equivalents, applied to real-world datasets.

git clone https://github.com/the-mrinal/ML-Notebook.git

---

Curriculum

Foundations

Module	Topic	Notebooks
02	NumPy Arrays	Array operations, histograms, visualization
03	Graphs	Plotting fundamentals
04	Plotting Graphs	Advanced matplotlib techniques
05	Intro to Datasets	Exploratory analysis on the Titanic dataset

Regression

Module	Topic	Highlights
06	Linear Regression	From-scratch fit() & predict(), R² score, cost functions
07	Multivariable Regression	Multiple features, gradient descent (iterative improvements)
08	Combined Cycle Power Plant	Feature-based regression with gradient descent

Classification

Module	Topic	Highlights
09	Logistic Regression	Binary classification, Titanic survival prediction
10	Confusion Metrics	Precision, recall, F1-score, classification reports
11	Decision Tree	sklearn + graphviz visualization, Iris & Titanic datasets
15	Naive Bayes	From-scratch with Laplace smoothing, Iris dataset
16	KNN	From-scratch + sklearn, cross-validation, hyperparameter tuning
18	SVM	Linear SVM, decision boundary visualization

Feature Engineering & Dimensionality Reduction

Module	Topic	Highlights
12	Feature Scaling	StandardScaler, MinMaxScaler on Wine dataset (13 features)
13	Gradient Descent Projects	Boston housing, logistic regression with gradient descent
19	PCA	Eigendecomposition, 3D-to-2D, face images (LFW), breast cancer data

Deep Learning

Module	Topic	Highlights
24	Neural Networks (TensorFlow)	3-layer NN on MNIST — 96% accuracy
24	Neural Networks (sklearn)	MLPClassifier on various datasets

Projects

Module	Project	Dataset	Approach
17	Text Classification	20 Newsgroups (19,997 docs)	Multinomial Naive Bayes, NLP pipeline
20	Used Car Price Prediction	297,222 records	Feature engineering, regression
21	CIFAR-10 Image Classification	60,000 images	PCA + Random Forest / LR / SVM / KNN comparison
22	NLP	Movie reviews	Stemming, lemmatization, preprocessing

---

Datasets Used

Dataset	Type	Size	Used In
Titanic	Classification	891 records	Modules 05, 09, 11
Boston Housing	Regression	506 records	Modules 06, 13
Iris	Classification	150 samples	Modules 11, 15, 18
Breast Cancer	Classification	569 samples	Modules 16, 19
Wine	Classification	178 samples	Module 12
CCPP	Regression	9,568 records	Module 08
Diabetes	Regression	442 records	Module 06
20 Newsgroups	Text Classification	19,997 docs	Module 17
Used Cars	Regression	297,222 records	Module 20
CIFAR-10	Image Classification	60,000 images	Module 21
MNIST	Image Classification	70,000 images	Module 24
LFW Faces	Dimensionality Reduction	13,000+ images	Module 19

---

Tech Stack

Category	Libraries
Data Processing	pandas, numpy
Machine Learning	scikit-learn
Deep Learning	TensorFlow
Visualization	matplotlib
NLP	nltk
Tree Visualization	pydotplus, graphviz

---

Getting Started

# Clone the repository
git clone https://github.com/the-mrinal/ML-Notebook.git
cd ML-Notebook

# Install dependencies
pip install numpy pandas matplotlib scikit-learn tensorflow nltk pydotplus

# Launch Jupyter
jupyter notebook

Navigate to any numbered module folder and open the .ipynb files to begin.

---

Repository Structure

ML-Notebook/
├── 00. data/                        # All datasets (CSV, images)
├── 02. NumpyArrays/                 # NumPy fundamentals
├── 03-04. Graphs & Plotting/        # Matplotlib
├── 05. Intro to Datasets/           # EDA on Titanic
├── 06. LinearRegression/            # 5 notebooks
├── 07. MultivariableRegression/     # 6 notebooks
├── 08. CombinedCycle/               # Regression project
├── 09. Logistic Regression/         # Binary classification
├── 10. ConfusionMetrics/            # Evaluation metrics
├── 11. DecisionTree Implementation/ # 5 notebooks
├── 12. featureScaling/              # Normalization techniques
├── 13. GradientDescentProjects/     # 5 notebooks
├── 15. NaiveBayes/                  # From-scratch + sklearn
├── 16. KNN Implementation/          # From-scratch + sklearn
├── 17. textClassificationProject/   # Full NLP pipeline
├── 18. SVM/                         # Support Vector Machines
├── 19. PCA/                         # Dimensionality reduction
├── 20. UsedCarProject/              # Real-world regression
├── 21. cifar10/                     # Image classification
├── 22. NLP/                         # Text preprocessing
├── 24. Neural Network/              # TensorFlow & sklearn NN
├── LICENSE
└── CODE_OF_CONDUCT.md

---

License

MIT License - Copyright (c) 2018 Mrinal Chandra

---

Topics

machine-learning deep-learning python jupyter-notebook scikit-learn tensorflow data-science neural-networks classification regression nlp pca naive-bayes knn svm decision-tree gradient-descent mnist cifar10