The Student Room Group

Computer science project

I am in year 12 and I was asked to do an experiment (don't need to be super difficult) for computer science, I want to do something related to machine learning, any suggestions on what can be experimented? Thanks
Reply 1
In terms of for the coursework project? My OCR A-Level CS project uses LSTM machine learning, I'd recommend looking at what application of ML interests you most and would suggest using premade libraries to achieve this. I've found fully grasping machine learning to be quite challenging, libraries will streamline this process and make it far less difficult.

There's so many fun applications, you could create a project that attempts to predict what object is being shown to the camera or use ML to outline a persons 'hitbox' or hand, to name a few. My project attempts to predict stock market trends - albeit somewhat unsuccessfully, at the end of the day though your coursework doesn't actually have to succeed to achieve full marks so go with something you think will keep you the most focused throughout the duration of the project even if it may be unsuccessful. Generally ML is best applied for pattern recognition, pick something where patterns are a main focus as a rudimentary model should be able to predict patterns with good success.
Developing a simple model to differentiate between categories based on the input is a good starter project! A good example of this is the simple spam filter used on email services, you could definitely implement something like this yourself for a start. 😊

-- Elliot
Reply 3
Original post by Bolognase
In terms of for the coursework project? My OCR A-Level CS project uses LSTM machine learning, I'd recommend looking at what application of ML interests you most and would suggest using premade libraries to achieve this. I've found fully grasping machine learning to be quite challenging, libraries will streamline this process and make it far less difficult.

There's so many fun applications, you could create a project that attempts to predict what object is being shown to the camera or use ML to outline a persons 'hitbox' or hand, to name a few. My project attempts to predict stock market trends - albeit somewhat unsuccessfully, at the end of the day though your coursework doesn't actually have to succeed to achieve full marks so go with something you think will keep you the most focused throughout the duration of the project even if it may be unsuccessful. Generally ML is best applied for pattern recognition, pick something where patterns are a main focus as a rudimentary model should be able to predict patterns with good success.
Lol, we've done the exact same thing πŸ˜‚.... a Market value predictor for a stock. Mine also checks whether there has been a change in the current stock price and if so, it sends out an email to all users on a mailing list informing them that there has been a change in their stock value. It checks every half hour for the stock price and it doesn't inform the users unless there has been a decent drop. I've tested mine and it works well.

Quick Reply

Latest

Trending

Trending