gdsc-igdtuw-autumn-of-code-2022 / ml-projects-2023 Goto Github PK

This Project Predicts whether a mushroom is edible or poisonous.

Based on this Kaggle dataset the following has been performed:
-Elementary Data Analysis: Data visualization and Statistical Analysis
-ML model training and testing: Data cleaning, preprocessing, and ML algo/model.

Accuracies achieved: Train accuracy - 100%
Test accuracy- 96.64%
Some snaps are as follows:

I'll be working with data for 768 patients of Pima Indian Women and drawing out some conclusions after data visualization using graphs and plots.

In this data 65% womens are healthy whereas 35% women are diabetic .
From this data we'll be able to infer that :
1.Women after age of 30 are more prone to diabetes.
2. Women after more pregnancies there are more chances to have high glucose levels.
3. With age the blood pressure increases.
4. Glucose is the most important factor in determining the onset of diabetes followed by BMI and Age.
5. Other factors such as Diabetes Pedigree Function, Pregnancies, Blood Pressure, Skin Thickness and Insulin also contributes to the prediction.
6. Diabetic people tend to have higher BP which leads to risk of many other diseases such as hypertension, heart diseases,strokes etc.
7. Obese people develop more chances of higher insulin levels which in turn lead to pre diabetic or diabetic conditions.

I used media pipe, pyautogui, and cv2 in python (spyder) for this project 
—->) MediaPipe provides open-source, cross-platform, and configurable machine-learning solutions for live and streaming video.
such as human posture, facial tracking, and so on

—->)PyAutoGUI is a cross-platform Python GUI automation package for humans. Used to control the mouse and keyboard programmatically.
PyAutoGUI currently only works on the primary display. PyAutoGUI is unreliable for the screen of a second monitor (the mouse functionalities on multi-monitor setups may or may not operate depending on your operating system and version).
These are the most crucial modules in this project.

NOTE: This project is highly beneficial for those who do not have hands and can only handle the mouse with their eyes.

please check the link

---->https://drive.google.com/drive/folders/1ybaQoX74W9uJsg1NvkJxx_GUVzI6s3eP
here is some proof of the project. In the video the mouse is controlled by the eyes of the user.

Data visualization was done along with pruning of the decision tree to get a more optimized result.
Prediction results as follows:-

As the name suggest I've built a SVC classifier that takes an image as an input and outputs 0 or 1(tomato or apple), based on the resemblance of the image with the two.

This project is used to detect drowsiness. In this Project, we check how long a person's eyes aren't blinking. If the eyes have been closed for an extended period i.e. beyond a particular threshold value, the program will alert the user by saying
"wake up"