Financial Forecasting Using Linear Programming

In the stock market, investors suffer from a financial crisis due to defective forecasts. Accuracy has always been a challenge in predictions with pure statistical and ML models. Movement of stock market is contingent with political stability and other exogenous variables also. No such application h

Project Title

Financial Forecasting Using Linear Programming

Project Area of Specialization

Artificial Intelligence

Project Summary

In the stock market, investors suffer from a financial crisis due to defective forecasts. Accuracy has always been a challenge in predictions with pure statistical and ML models. Movement of stock market is contingent with political stability and other exogenous variables also. No such application has been designed which helps better stock predictions along with the
impact of local and global events.

We are aiming to develop an open web application which will provide the accurate forecasts of the stock market with hybrid model having both Machine Learning and statistical algorithms in a coherent architecture. We will also provide a decision support system to help the investors guide based on buy-hold-sell strategy.

Project Objectives

• To provide Decision Support System
• To provide the Summary of Stock
• To provide Accurate Forecasts
• To provide Live Forecasts
• To provide study on imapct of exogenous variables on stock
   

Project Implementation Method

Our implementation is divided into following three major parts:
•    Data Preprocessing 
•    Forecasting Model
•    Data postprocessing

In data preprocessing component the main goal is to make time-series data stationary and this can be done by removing linear trend from series. To achieve our goal, we are using Linear Regression Technique which calculates linear trend line and then removes it from series data.

Second component is our forecasting model. We are using a hierarchical model which is a combination of Statistical and Machine Learning Techniques known as “ES-RNN” which stands for Exponential Smoothing and Recurrent Neural Network. This method was purposed by Slawek Smyl in M4-Competition and got the award of best forecasting model. This model supports cross-learning which enables our model to learn from multiple time series at same time. Let’s dive into further details of our model. It has major three parts, in first part this model use Holt-Winters Exponential Formulas to deseasonalize and normalize data. Neural networks do not perform well on seasonal data so it's essential to first remove seasonality from time-series data. The second part is Recurrent Neural Network which actually makes forecasts. In this model Dilated LSTM-based stacks are used which is a form of RNNs. This stack has two block of LSTMs and each block have two layers followed by linear adapter layer whose work is to make the output of hidden layers adaptable for output layer. Third part is known as reforming data part in which model reintroduces seasonality in predicted values.

Third component of our model is data postprocessing in which we add the removed trend in the predicted values abg generate or finals forecats.

Benefits of the Project

This project comes under the umbrella of FinTech. The total market capitalization of all publically traded securities in world is almost US$97.3 trillion as per 2020 data. 

We are addressing this huge world of investment where we are helping the investors in mkin decisions on whether should the buy more stock or hold for the time being. It will saves millions of dollars. Forecasts will help the investors in analyzing the whole pattern of market and it will influence their decisions.

Technical Details of Final Deliverable

We will have:

• Graphical Summry of Stock
• Hybrid model of Exponential Smootihng and LSTMs generating the forecasts.and its graphical implementation
• Forecasts generated will be live based on the real time data obtained by API calls.
• Decision support system based on buy-hold-sell strategy.

Final Deliverable of the Project

Software System

Core Industry

Finance

Other Industries

IT

Core Technology

Artificial Intelligence(AI)

Other Technologies

Big Data

Sustainable Development Goals

Decent Work and Economic Growth, Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com