Data is the key to machine learning, in mlTradeTool huge amounts of data will be processed, among which are: Stock market, currencies, futures, options, commodities News, Wikipedia, World and country statistical information, Google trends, Archive.org, Google analytics, and any other that can provide relevant information.

To do all this, an important technical infrastructure will have to be deployed that downloads and process all new information
Then a machine learning system has to update predictions acording to the new data incorporated.

What is certain is that the world is interrelated, countries, currencies and machine learning is extraordinary to find patterns. In MLTradeTool predictions will be:

Intraday
Next day
At 2 days
At 7 days
At 1 month
At 3 Months
At 6 Months

Here the investor can see his investment portfolio and can transfer funds from one account to another. Integration with some brokers will be necessary for this.

This module will have the necessary tools to do manual and automatic trading based on prediction rules.

Alerts will keep the relationship with customers loyal. They will be configurable but will generate a minimum of iterations with the clients. They will be of two types, informative to notify the client of an event, such as a closing of operations and alerts of investment or sale opportunities

There are various techniques that have been used so far to predict the evolution of the stock market like technical analysis, fundamental analysis.
Machine learning can add value to both by finding patterns of what one specific event may affect to another.

The stock markets are subject to a number of variables that cannot be modeled, the model will be better with more variables, but it is impossible to collect all the variables because at least one of them is unpredictability: human stupidity.

Machine Learning algorithms typically work reasonably well under known and proven circumstances, but they often fail when the circumstances become different than trained.

For example, when the price of oil falls, airline companies rise because they earn in profitability, when the price of oil rises they usually have problems. This relationship, however, at the time of Covid-19 was false, the oil dropped a lot but the airlines did not benefit, in fact it has been a black period for them. At this point, any model for airlines will fail unless it includes parameters such as mobility of people, number of flights, average price, etc.
From the same point of view, when the stock market rises or falls, most stocks tend to behave like the market. But again in exceptional periods, such as that of the Covid-19, a model could learn that if the market falls, it is an indicator that pharmaceutical companies will rise. Machine learning will perform much worse than an expert on new not seen before new circumstances ... or perhaps not if the model is good enough and large
S&P index fell sharply in March 2020, no expert could have predicted, but probably a machine learning system could have taken into account the data that was already known about what was happening in China.

With PRModel we can model huge data volumes at high speed and achieve high prediction reliability. In the stock market a predictor is considered good if it is correct at 51%, the issue is that we do not need to predict whether a value will rise or fall, but which of the important predictions are likely to be correct. The implementation will not only predict, but will also estimate the probability of success based on the predicted asset. For this reason, we believe that the model will achieve very high success rates on a sustained basis, at the cost of not predicting doubtfull cases.