To achieve quality outcomes, machine learning requires data, large amounts of data. That is why we are developing own universal crawlers used to get the necessary content.
We have practically solved all issues such as limiting access by captcha codes, chanching website structures and practical problems concerning the identification of correct information in the text. For example, when downloading reviews, the same items are commonly tagged differently and, for aggregation, it is necessary to identify them as the same product.
Data Harvesting – Extract, Transform, Loadevent_note 11.10.2021 person Tereza Zemanová
A good machine learning model needs rich and wide datasets.
Complexity of each calculation in the StockPicking Labevent_note 24.08.2021 person Maša Vodalov
When calculating the results and preparing a list of undervalued and overvalued stocks in our StockPicking Lab, for every asset, stock, and ticker we use 2237 features.
New SFA featuresevent_note 28.07.2021 person Maša Vodalov
We have deployed a new version of our SFA (Summary of Financial Articles) software.
Model Diagnostics with Learning Curvesevent_note 22.07.2021 person Maša Vodalov
Learning curves can bring important insight during the design process of a machine learning model.
Feature Selectionevent_note 28.06.2021 person Tereza Zemanová
We have been researching the selection of variables enterings. When there are too many variables, the model has a worse ability to generalize and is more prone to errors.
New version of the opinio applicationevent_note 20.06.2021 person admin
We have released a new version of the opinion app! The app underwent a complete redesign, we focused on a better UI and added a large number of products.