$tarTrader

Inspiration

We knew how EVE Online had such a large player based economy and wanted to extend machine learning to economics and gaming.

What it does

It takes historical market data in the game EVE Online and tries to predict the next day's price based on it.

How we built it

We used a long short-term memory model (a kind of recurrent neural network), trained it on the historical prices of EVE Online items, and predict tomorrow's price.

Challenges we ran into

We were having trouble descaling the data for predictions because of the specific dimensional requirements for the LSTM

Accomplishments that we're proud of

We are proud of creating a tool that can be used not only in the game, but can be extended to real life commodities exchange

What we learned

We learned how to use the ESI, apply LSTMS to data, and use TA-Lib (for TA).

What's next for $tarTrader

We can improve it by adding things like volume thresholds, to improve accuracy

Built With

• esi
• python
• scikit-learn
• ta-lib
• tensorflow