Is possible to train a ML in google collab to read charts and output the % it will bull or bear?
>>60807075 (OP)
Yes, this was literally a grad school CS project our team chose to pursue recently. Use TenserFlow and LSTM modeling to tune the hyperparameters on historical TA data/methods. Unfortunately, even with substantial tuning, our bot was still wrong more than 50% of the time because nothing in crypto makes sense, effectively making it useless to turn a profit. Rather, broader crypto is a hype cycle, with specific coins having their own micro hype cycles, so if you can quantify that somehow maybe you could make a better ML bot
Unfortuantely, you're likely a low IQ midwit having made this post so you're probably better off not trying to use your smooth brain for intensive high IQ workloads
Yes, this was literally a grad school CS project our team chose to pursue recently. Use TenserFlow and LSTM modeling to tune the hyperparameters on historical TA data/methods. Unfortunately, even with substantial tuning, our bot was still wrong more than 50% of the time because nothing in crypto makes sense, effectively making it useless to turn a profit. Rather, broader crypto is a hype cycle, with specific coins having their own micro hype cycles, so if you can quantify that somehow maybe you could make a better ML bot
Unfortuantely, you're likely a low IQ midwit having made this post so you're probably better off not trying to use your smooth brain for intensive high IQ workloads