Parallel Stock Prediction using Python

By SUMIT HARESH GANGARAMANI

Argument that real time stock prediction can be done effectively using parallel execution has been presented.(Using High-performance computing and combining the outputs of all the models.
).

Introduction

Large availability of stock market data but it’s use in forecasting stock prices is less.

Motivation arises from the above fact to apply model parallelism.

Argument that real time stock prediction can be done effectively using parallel execution has been presented.

This happens by combining the outputs of all the models.

 

Objectives

Model parallelism

Data exploration

Stacking Models

Multiprocessing

 

Methodology

Different Models used are LSTM, GRU, Vanilla, CNN-Seq2seq:

• LSTM 
• Vanilla
• CNN-Seq2seq

Different libraries used are:

• Multiprocessing
• Tensorflow
• Keras
• Others