https://github.com/ganesh2409/strive_towards_ai

This repository contains materials from a two-session workshop on Machine Learning and Deep Learning. Session 1 covers data preprocessing techniques including data cleaning, feature engineering, and exploratory data analysis. Session 2 focuses on building and training a neural network using TensorFlow and the Fashion MNIST dataset.
https://github.com/ganesh2409/strive_towards_ai

data-cleaning-and-preprocessing deep-learning exploratory-data-analysis machine-learning

Last synced: 11 days ago
JSON representation

This repository contains materials from a two-session workshop on Machine Learning and Deep Learning. Session 1 covers data preprocessing techniques including data cleaning, feature engineering, and exploratory data analysis. Session 2 focuses on building and training a neural network using TensorFlow and the Fashion MNIST dataset.

• Host: GitHub
• URL: https://github.com/ganesh2409/strive_towards_ai
• Owner: Ganesh2409
• Created: 2024-01-21T17:42:43.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2024-08-10T10:35:24.000Z (almost 2 years ago)
• Last Synced: 2025-02-28T11:51:29.004Z (over 1 year ago)
• Topics: data-cleaning-and-preprocessing, deep-learning, exploratory-data-analysis, machine-learning
• Language: Jupyter Notebook
• Homepage:
• Size: 43.6 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
## Session 1: Data Cleaning, Feature Engineering, and EDA

In the first session, participants dived deep into the essentials of preparing data for machine learning models. The session was structured into the following key topics:

1. **Data Cleaning**: 

   - Handling missing values through different techniques like dropping rows/columns and imputing with mean values.

   - Dealing with outliers and understanding their impact on model performance.

2. **Exploratory Data Analysis (EDA)**:

   - Visualizing data distributions and relationships between features using various plotting techniques like box plots, scatter plots, and KDE plots.

3. **Feature Engineering**:

   - **Feature Construction**: Creating new features using domain knowledge.

   - **Feature Encoding**: Applying techniques such as One-Hot Encoding, Ordinal Encoding, and Label Encoding to convert categorical data into numerical values.

   - **Feature Scaling**: Understanding and applying scaling techniques like Standardization and Normalization.

4. **Feature Transformation**:

   - Transforming features using techniques like log, square root, and reciprocal transforms to make data more suitable for modeling.

**Presentation**: A detailed presentation for this session is available, providing step-by-step guidance and insights.

---

### Session 2: Building a Neural Network with TensorFlow

In the second session, participants got hands-on experience with deep learning. This session covered the following key topics:

1. **Introduction to TensorFlow**:

   - Setting up and importing necessary libraries.

   - Loading and preparing the Fashion MNIST dataset for model training.

2. **Building a Neural Network**:

   - Constructing a simple sequential neural network using TensorFlow and Keras.

   - Understanding the architecture with a dense hidden layer and an output layer with softmax activation for multi-class classification.

3. **Training and Evaluating the Model**:

   - Training the model on the training dataset and evaluating its performance on the test dataset.

   - Using the model to make predictions on new, unseen data.

4. **Visualization**:

   - Visualizing sample images from the dataset and understanding model predictions.

**Presentation**: The session's presentation includes step-by-step instructions and explanations, making it easier to follow along and understand the concepts.

---

### Additional Resources

Both sessions include PowerPoint presentations that provide visual aids and additional explanations. These presentations are invaluable for anyone looking to reinforce their understanding or revisit specific topics covered during the sessions.

### Final Thoughts

In this fast-evolving world of technology, the skills learned here will equip you to tackle various machine learning and deep learning problems. Remember, as long as you can convert your data into numerical form, the possibilities with ML/DL are endless!