StockML-TF

Tensorflow JS implementation under NodeJS for server side applications, with simple HTTP API for remote control

Install:

- npm install

HTTP API

• GET http://localhost:3000/model_name/shapes // Input and Output tensor shapes for training and Input useable for prediction
• GET http://localhost:3000/model_name/status // Loaded,Training,Ready when it is Ready your Model is ready for accept prediction
• POST http://localhost:3000/model_name/predict + {input: your_data} // Accept data for prediction/evaluation and response the result

Available modells:

• Rnn
• LSTM
• LSTM with Cells

What we are looking for:

• Smma cross(SAwesome oscillator) indicated trades predict with classification method Good/Bad (An avarage good trade on the picture below)

Results:

• AO,Trix,Mome -> lstm_hidden_cells with 3x50 Epoch: loss=0.0812 val_loss=0.0822
• AO,Trix,Mome,RSI -> lstm_hidden_cells with 3x50 Epoch: loss=0.0817 val_loss=0.0742

Example data structure:

• Useable for utils.js -> trade_singal_extractor(name, indicator_count)

• trade_history_ao_mome_trix, 3 // indicator count = 3

• trade_history_ao_mome_trix_rsi, 4 // indicator count = 4

{""buy_price":0.000009315, // Price at buy
"sell_price":0.0000093375, // Price at sell
"buy_in":[-0.03315335485388681,-0.003207981457875597,-0.00002304744910475165,-0.02536090073326459,-0.018446781036718374,0.00021997229895376336,-0.013662891592516064,-0.04892210958552523,0.0003945799145643747,-0.0054359002795636375,-0.12742222435927486,0.0005105833430496531,0.005674614807288707,-0.15946790608221129,0.000550478723893022,0.009904300453449123,-0.07858798245402587,0.000545215674541133], 
// History of the Awesome Oscilator, Momentum indicator and Trix oscilator values 5 x N
"time_history":[[-0.0010844294644542386,-0.12262857967250129,-0.002553803934646215,0.0311535],[0.002191581194322352,-0.1458578767260521,-0.0024641477721742966,0.03114925],[0.0127033532331586,-0.17712307248420878,-0.0024202076649104676,0.03113775],[0.014649884250143663,-0.1386829131428291,-0.0023879032743034934,0.031142999999999997],[0.011957486321610025,-0.12985763755289925,-0.0023453057852007575,0.0311475],[0.00938266583288114,-0.10182809493265443,-0.002291109663488111,0.03114825],[0.004153614790243935,-0.0785829410868466,-0.0022107647248970075,0.03115275],[0.0033131312004486936,-0.07137873234578374,-0.0021237582055966264,0.031149500000000004],[-0.006662629207976794,-0.013636545943118694,-0.001985401976833945,0.031162000000000002],[-0.007372564667509738,0.0160512355438458,-0.0018374887635785875,0.03115525],[0.004433368222877852,-0.028894079121610794,-0.0017450508455585378,0.03113925],[0.0036896541986646113,-0.03850905765127033,-0.0016548037429453544,0.0311495],[0.0004994915597795197,-0.005617120984757751,-0.0015538830058337005,0.031153],[-0.00030753452845337393,-0.04091226325037001,-0.0014522383138704242,0.0311515],[-0.013320563391349323,-0.024858865794208374,-0.0012888176303293974,0.03116825],[-0.016569947167483202,-0.06253557713122859,-0.001105282299140864,0.03116275],[-0.023762793873273454,-0.030468248877479898,-0.0008827080264728338,0.0311705],[-0.03382938080333411,0.028873685645747216,-0.0006052615863505473,0.03117925],[-0.03650411762233365,-0.020844858134716482,-0.0003076104684529568,0.03117625],[-0.03315335485388681,-0.003207981457875597,-0.00002304744910475165,0.031171249999999998],[-0.02536090073326459,-0.018446781036718374,0.00021997229895376336,0.031164999999999998],[-0.013662891592516064,-0.04892210958552523,0.0003945799145643747,0.031156749999999997],[-0.0054359002795636375,-0.12742222435927486,0.0005105833430496531,0.03115575],[0.005674614807288707,-0.15946790608221129,0.000550478723893022,0.03114775],[0.009904300453449123,-0.07858798245402587,0.000545215674541133,0.03115075],[0.008927634413868808,-0.016046213093701134,0.0005230442303938432,0.031155000000000002],[0.007108393110419857,0.0224724511826091,0.0004935165869079938,0.03115625],[0.005031041914643647,0.004814520594109695,0.00046280059107228766,0.03115725],[0.005302450683589745,-0.012837382457719926,0.0004232199345421961,0.031155],[-0.005007432929114201,0.028080872913993637,0.0004294938223514677,0.03116875]] // Full time frame history from Buy to Sell Indicators + actual price N x 4
}