Home - Deep Learning for Coders, Traders & Others Interested

On Code2Trade we focus on (Artificial) Intelligence in general but especially on its usage in Financial applications

To be able to understand the behavior of markets we use Models. For models we need Data: the best predictor of future behavior is past behavior! We also use knowledge. Technical Analysis teaches us the workings of markets. Last but not least, we apply our Coding skills to knit things together and bring our algorithms to live.

On all these topics you’ll find posts with reusable, working code examples.

Recently we decided to make room for another interesting topic. Photo Intelligence. Making use of smart algorithms in managing photo collections. For instance: face and body recognition to automatically detect specific persons in images, smart hashing techniques to search and optionally remove low resolution duplicates.

Be sure to come visit regularly to discover posts on this topic!

In the last couple month we’re doing a series on Price Patterns and Algotrading. Things started with the empiric discovery of weekly price patterns in crypto currencies, especially with the so called meme coins. The basics are described in this post on Weekly Price Patterns in Meme Coins and Traditional Cryptocurrencies. The whole idea behind Algotrading is using knowledge to automate trading strategies. In Automate Your Edge: A DIY Crypto Signal Bot with Python and Windows Tools we demonstrate the use of pattern based knowledge to create an entry- & exit-signals generating bot, you can easily host on your own Windows machine.

This concept is extended in the next post, by exploring whether predictable intra-week patterns can be used for algorithmic trading. In From Pattern Recognition to Signal Execution: A System for Weekly Cycle Trading in Crypto Assets we developed a forward-looking classifier capable of forecasting whether the upcoming week is likely to follow the expected pattern. Having established robust standard weekly price patterns, we can confidently use these as reference profiles to verify the factual early-week resemblance to this reference using real-time data. We developed a script that script efficiently processes a watch-list of assets and identifies suitable candidates for a trading signal generation bot. The architecture thus combines early-week pattern recognition with intraweek signal generation, creating a layered decision system.

After month of using the setup described above, we  evaluate things in a post called Pattern Prediction and Recognition in Meme Coin Time Series. For Pattern prediction and recognition we propose a layered framework combining standard weekly reference profiles, forward-looking classifiers, and entropy-aware real-time resemblance checks to evaluate meme coin price behavior. This approach seeks to balance predictive modeling with live confirmation, making algorithmic trading more adaptive to volatile markets.