Aim of the experiment
We want to make available to the small investor, Machine Learning techniques to be able to identify what are the best times to buy shares on some Stock Markets.
Specifically we make predictions of purchase with the goal of winning 10% in the 20 next coming sessions, if it does not happens our strategy is to sell that 20th day and assume the gain/loss accruing at the closing price of that day.
By now, we model and predict only some stocks from Mercado Continuo (Spanish stock market), Mercado de Valores (Argentina stock market), NASDAQ and Oslo Børs (Oslo Stock exhange from Norway).
You can check for the results of the predictions done that are in the period of 20 coming sessions and for the older ones also.
What is Machine Learning?
Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it learn for themselves.
The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data and make better decisions in the future based on the examples that we provide. The primary aim is to allow the computers learn automatically without human intervention or assistance and adjust actions accordingly.
Models used to predict
The origin of the data is a time serie with the historical data of each one of the stocks from 2012. For each session we have the quotation in the opening, the maximum and the minimum of the day and the volume.
Starting from a time series, we can not define a model directly with the observations we have, since the observations are not independent of each other, but depend temporarily on each other.
To deal with these time series of financial data, we have calculated the values of different technical indicators that are commonly used by stock market analysts, and we have done some transformations.
Every day, we compute 3 classification models for every stock, and after modeling them, we choose the model that fits better for each one.
When the market closes, or in some markets twice a day, we predict with the best model we have for that stock
Machine learning algorithms we use to compute the models are Logistic Regression (LR), Support vector machine (SVM) and Random forest (RF)
Atention, legal notice
Past Performance is No Guarantee of Future Results
The information contained herein and the one resulting from the execution of the prediction system we have constructed is presented for informational purposes only and does not constitute an investment recommendation, nor invitation, offer, request or obligation on the part of BrokerTrade, to carry out operation or any transaction.
The dreams of yesterday are the hopes of today and the reality of tomorrow.