SubSeaSignal

Inspiration

The main inspiration for this project came from a speech by a Navy official in a seminar that one of our team members attended. He stated that declogging the sonar data for the general officers in the navy is trickier because of their lack of knowledge in sonar signal reading.

What it does

The ML model takes the 3-dimensional input of the detected object in the ocean and using some complex mathematical calculations, it predicts the object where it is a rock or a mine and also can plot a graph of all the objects against each other.

How we built it

We took the dataset from UCI archives to train the model and test it. We used different statistical measurements to get accurate values to classify them into groups.

Challenges we ran into

The dataset is very confusing to understand and also to work with. It was really challenging to improve accuracy.

Accomplishments that we're proud of

We are proud that we were able to accomplish the task we set out for and also try to help the needed government officials by improving the model using much more data.

What we learned

We have learned that it was necessary to pre-process your data to really understand which attribute is to be considered that will help to bring more accuracy

What's next for SubSeaSignal

We can improve our model using more data as this kind of data is not publicly available in abundance. We can also try giving this to the government and see what they can do with it.