Machine Learning Stock Prices

Machine Learning Stock Prices

Theme Automation & Big Data

challenge

Prices of listed securities are subject to changes every second and in some cases to strong fluctuations. In order to minimise losses, it should be possible to predict whether these will remain stable in the coming minutes or whether they will be sufficient to exceed the limits.

target

A portfolio manager or day trader should be automatically given a rating as to whether certain values in his portfolio are threatened by strong outliers in the short term. The manager can then react in a targeted manner and, if necessary, initiate countermeasures to compensate losses.

Result

We have designed and combined various modern machine learning algorithms and made them available via an API. In addition, we have created a plugin which the portfolio manager can integrate into his analysis charts and which is in contact with the API. By specifying certain limit values and continuously analyzing the transmitted value fluctuations, an alarm is sent to the portfolio manager at an early stage that expresses the probability that a certain investment will exceed or fall short of the given limit values. The calculation of the probability of such a limit being exceeded is also visualized in charts.