Classmatch

• 

  The title page of our slideshow

Inspiration

As high school students affected by the Coronavirus pandemic, we know how difficult it is to experience efficient learning time, especially when you're stuck in your house all day. It's terrible when teachers schedule meetings at 4 or 5 PM when you're exhausted from schoolwork and your sole desire is to take a nap or go outside. What's even worse, though, is when they schedule meetings at 7 AM when you just want to go back to sleep and continue your dreams. That's why we created Classmatch!

What it does

Classmatch is a machine-learning based hack which determines the time of day at which the student is most productive.

To do this, it will allow the user to either complete a series of questions which we drafted or automatically track certain attributes of their daily schedule for ten days.

The machine learning model will then analyze this data to determine the user's most productive time and place them in one of four categories (Productive in the Morning, Productive in the Afternoon, Productive in the Evening, and Productive at Night). Today, we have created two attributes, the sleep time and the waking up time. In addition, we have only created Morning and Not Morning as the binary options for training the model later in the program.

(In this project, we have used synthetic data for a large number of students and then uses that data to train a Machine learning model).

How I built it

We leveraged Google's Colaboratory platform to develop our project. The project is built in Python and a Machine Learning library called - scikit learn.

First, we decided what factors we wanted to include in the model. Then, we proceeded to generate synthetic data. Following that, we split the data in to training, which the majority of the data fell under, and testing, which a small portion of the data fell under. Then, we scaled it to make functions faster to perform for the computer. We trained it and tested its accuracy. As the last step, we decided to include user input.

Challenges I ran into

A challenge we ran into was connecting the front-end code to the back-end code, so we ultimately decided to put that on hold in favor of writing a good, robust model.

Accomplishments that I'm proud of

We are proud of the fact we were able to develop and train a Machine Learning Model using learning Logistic regression considering how complex and difficult machine learning and the general field of artificial intelligence can be.

We are also very proud that we were able to come up with a minimum viable product on time as planned.

What I learned

We have been hearing about machine learning and artificial intelligence many times. It was very exciting to learn this field by doing and building something of our own. It gave us a new perspective of why so many people are fascinated with this technology.

This hack contained a lot of learning experience for us because when we first started out, machine learning was a hazy concept but we learned what scaling data meant, how to train and test model, and how to verify its accuracy using simple concepts.

What's next for Classmatch

Next, we plan to expand on our model by adding more factors (eating times, social interaction times, and working times) and connecting the front-end code to the back-end code. We also want to make our program accessible to teachers and students and connect a teacher with a group of students based on the times when they are most productive.

We hope that this concept will go a long way in ensuring enjoyable and much effective learning experiences for students.

PLEASE NOTE: In addition to the GitHub repository, we have also included the full code that can be tried out in Google Colab Enviroment. https://research.google.com/colaboratory/faq.html

Built With

• colab
• google
• python