• New SFA features

    New SFA features
    We have deployed a new version of our SFA (Summary of Financial Articles) software. We have recently added: filtering articles according to the date of their publication, from one day to an unlimited long history. determining sentiment either directly in the articles about selected tickers or in all articles where…

    Read more

  • Model Diagnostics with Learning Curves

    Model Diagnostics with Learning Curves
    Reviewing learning curves of models during training and plots of the measured performance can be used to diagnose problems with learning, such as an underfitting or overfitting model. Besides that, it can also be used for diagnosing whether the training and validation datasets are suitably representative and to observe generalization…

    Read more

  • Feature Selection

    Feature Selection
    As part of the research used, for example, in the StockPicker application we have been researching the selection of variables enterings. When there are too many variables, the model has a worse ability to generalize and therefore is less robust and more prone to errors. If you need to extract…

    Read more

  • Feature Binning and Quantile Transformation

    Feature Binning and Quantile Transformation
    As part of the research used, for example, in the StockPicker application, we have recently implemented the method Feature Binning and Quantile Transformation to better classify data. Due to upgraded data preparation, our machine learning models now achieve better results. You can see this in detail in the following lines.…

    Read more

  • Reporting of the StockPicking system 6/2020 – 5/2021

    Reporting of the StockPicking system 6/2020 – 5/2021
    It has been a year (03. 06. 2020 – 28. 05. 2021) already since we started trading with our “StockPicker” strategy on real Interactive Broker accounts. Therefore, we bring more detailed results, including graphs and comparisons with a benchmark, in our case the S&P 500 Index. First, let’s recall how our…

    Read more

  • Selection of significant predictors from a large number of time series

    Selection of significant predictors from a large number of time series
    Undervalued and overvalued stocks in our model are currently calculated as part of a proof of concept solution that is based on proprietary software with inner limitations. Therefore, a new robust production system is being prepared. In this way, we can take advantage of synergies with existing components from other…

    Read more

Further news can be found on the website ccfr.cz

Continue to the web