Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 4

DELHI TECHNOLOGICAL UNIVERSITY

SHAHBAD DAULATPUR, MAIN BAWANA


ROAD DELHI-110042

Machine Learning Project File


(SE-204)

Submitted By: Submitted To:


Ravi Kumar, Sameer Gupta Ms. Shweta Meena
Batch-A2 Assistant Professor
Roll No- 2K22/SE/135 & SE Department
2K22/SE/152
TOPIC: Predicting Calories Burned During Exercise
Using Machine Learning.

Problem Statement: Developing a machine


learning model to accurately predict calorie expenditure
during exercise sessions, utilizing factors like heart rate,
body temperature, and duration.

Sub Topic:
a. Data Collection:
How data was collected, mentioning the utilization of
two datasets containing exercise and calorie
information.
b. Data Preprocessing:
The steps taken to prepare the data, including merging
datasets, handling missing values, and converting
categorical variables.
c. Data Analysis:
The process of analyzing the data, focusing on
understanding feature distributions and identifying
correlations.
d. Model Training:
Detail the process of training the model, highlighting
the use of the XGBoost regressor.
e. Model Evaluation:
How the model's performance was assessed on test data,
emphasizing the use of mean absolute error.
f. Building a Predictive System:
How the trained model is used to predict calorie burn
for new individuals based on their exercise data.

Description:
Data Collection and Preprocessing:

Combine exercise and calorie datasets.


Handle missing values and categorical variables.

Feature Selection and Engineering:

Select features like heart rate, body temperature, and


exercise duration.
Employ feature engineering techniques for improved
performance.

Model Architecture:

Utilize the XGBoost regressor for its efficiency and


effectiveness in regression tasks.

Training Process:

Optimize hyperparameters and minimize loss functions.


Employ cross-validation techniques for robustness.

Evaluation Metrics:

Assess performance using metrics like mean absolute


error (MAE) or root mean squared error (RMSE).

Deployment and Utilization:


Integrate the model into fitness tracking applications or
use it as a standalone tool.

Continuous Improvement:

Refine and update the model based on feedback and


new data to ensure relevance and accuracy.

Feasibility:
Software Requirements:
The software tools/libraries used for development (e.g.,
Python, scikit-learn, TensorFlow).
Hardware Requirements:
Minimum hardware specifications for running the code
efficiently (e.g., CPU, RAM, GPU).

Programming Language:
We are choosing Python as the programming language.
The advantages of Python for ML tasks, such as its
extensive libraries and ease of syntax.

You might also like