Human Activity Recognition

➲ Project description

Human activity recognition is the problem of classifying sequences of data recorded by specialized harnesses or smart phones into known well-defined Human activities, The problem will be solved using K-Nearest-Neighbor (KNN) algorithm where we will train the model on dataset of collection of human activities in order to classifying it.

➲ Prerequisites

This is list of required packages and modules for the project to be installed :

• Python 3.x
• Pandas
• Numpy
• Seaborn
• Matplotlib
• Scikit-learn

Install all required packages :

 pip install -r requirements.txt

➲ The Dataset

Human activites dataset contain about 564 columns to describe human current activity
and a target column which describe what's the current state of a human.

Dataset features and target :

Dataset head :

➲ Coding Sections

In this part we will see the project code divided to sections as follows:

• Section 1 | Data Preprocessing :
  In this section we aim to do some operations on the dataset before training the model on it,
  processes like :

  • Loading dataset
  • Check missing values and dataset statistics
  • Visualize some dataset columns in order to have some insights for better understanding
  • Apply dimensionality reduction using PCA algorithm in order to transform the data from a high-dimensional space into a low-dimensional space
  • Scale the dataset
  • Split the dataset into training and testing sets
    

• Section 2 | Model Creation :
  The dataset is ready for training, so we create a K-nearest Neighbors "KNN" model using scikit-learn and thin fit it to the data.
  

• Section 3 | Model Evaluation :
  Finally we evaluate the model by getting accuracy, classification report and confusion matrix.

➲ Installation

1. Clone the repo

   git clone https://github.com/omaarelsherif/Human-Activity-Recognition-Using-Machine-Learning.git

2. Run the code from cmd

   python human_activity_recognation.py

➲ Output

Now let's see the project output after running :

Dataset head :

Dataset statistics :

Activites :

Accuracy and classification report :

Model confusion matrix :

➲ References

These links may help you to better understanding of the project idea and techniques used :

1. Human Activity Recognition research : https://bit.ly/384uASG
2. Models for Human Activity Recognition : https://bit.ly/3zhW57a
3. Standard Scaler : https://bit.ly/3mt2r04
4. Principal Component Analysis (PCA) : https://bit.ly/3knqZVD
5. K-Nearest Neighbors (KNN) : https://bit.ly/2XI2Ukq
6. Model evaluation : https://bit.ly/3B12VOO

➲ Contact

• E-mail : omaarelsherif@gmail.com
• LinkedIn : https://www.linkedin.com/in/omaarelsherif/
• Facebook : https://www.facebook.com/omaarelshereif