SUPERLAGOS PROJECT
Data Mining Team Proof of Work, October 2021
Data mining project is now one month old!
The aim of the project is to provide basic and advanced indicators to the Superalgos platform with reliable and clear data mines. Work has also to be done on extending the calculation capacities of Superalgos to implement high complexity level and computing resources demanding indicators.
Contributions to the project are expected to different forms and at all level of expertise and are opened to coders and non-coders:
- Code review
- Indicator coding
- Data science / mining expertise
- Trading expertise
- Codebase
- Documentation
The team is developing around the contributions of 3 main protagonists mostly working on the Data Mines and indicators.
Data Mine Contributions
Quantum Data Mine
Quantum8 has provided a brand-new Data Mine for his first contribution to the project. Quantum data mine comes with a standard Keltner Channel, the Moving Volume Weighted Average Price and to price regime identification tools: Choppy and Chop index.
VWAP shows the average trading price of an asset within the period based on both volume and price.
Choopy and Chop index are both designed to help to identify if the market is choppy (trading sideways) or not.
Omega-One Data Mine
Carl J is currently working on some major implementation that will take place during the beginning of November. Insights on this future contribution can be followed in the Data Mining Telegram channel.
Polus Data Mine
I introduced 4 new concepts in this Data Mine. First, I have started to implement indicators combinations with the Fisher MFI indicator, consisting in the Fisher transform of the Money Flow Index. The idea is to show here the capabilities and opportunities of using signal processing transform to identify underlying behaviors in indicators.
The Keltner channels are now available in two versions:
- Channels with Exponential Moving Average
- Channels with Mother of Adaptative Moving Average
The goal was here to have a standard Keltner channels indicator to initiate combinations with Bollinger Bands to explore volatility patterns. The use of the MAMA provides a more reactive version of the Keltner Bands, allowing a faster approach of pattern recognition.
A major progress has been done in both demonstration of Superalgos capabilities and signal analysis with the implementation of Monte-Carlo evaluation of the forecast on one future period of the price using the random walk theory.
Superalgos is now provided with a full set of volatility indicators like Relative Vigor Index, Volatility adjusted MA, Volatility bands… The volatility indicator bot also includes an original indicator specially developed and only available in Superalgos: the Weighted Volatility Oscillator provides an analysis of the position of the Keltner channels against the Bollinger Bands, pondered by the position of the price inside the Bollinger Bands.
Finally, the Signal-to-Noise ratio originally proposed by John Ehlers is now available. While providing a good leveling tool to identify trading opportunities, the benefits of this indicator reside also in bringing an implementation of the Hilbert Transform to be reused on other applications /combinations.
Update On the Background Tasks
The code of all the indicators provided in Superalgos up to October 2021 has been reviewed and each data mine tested to check the signal provided were stable. The next steps will be to provide documentation for each data mine not already provided with documentation. A reorganization of the data mines will have to be done. The target organization will be to group indicators by domain (Moving average, volatility, volumes, signal processing…). The build of data mines dedicated to trading strategies will be considered too.
Advanced indicators and indicators combination will have to be developed more and more to bring Superalgos to the state-of-the-art level in terms of indicators availability.
On the side of the computing capabilities the project still waits for skilled developers to join the party and bring new computing capabilities like multi-threading. Integration of signal processing libraries from C++ or Python will also be considered.
Finally, with the opening of the Superalgos Forum, the collective effort will be covered in the Data Mining section of the forum where development axis will be provided.
Conclusion
A lot of interesting work has been done in the Data Mining project this month on different level of the project:
- Two new data mines: Quantum and Polus
- Enrichment of Omega-One data mine under progress
- Indicators code review
The team is still under construction and we will need more support to fulfill our ambitious goals. Contributors from different horizons can join the team for different contributions:
- Indicator documentation
- Data Mine reorganization
- Indicator coding
- Codebase evolution
If you find this work interesting and useful for the project, feel free to put voting power on the claims of the actors of the project in the Governance system.
Want to join the team and put some claims in the Governance project to earn some SA Tokens?
Come and visit us in our regular channels!
- Superalgos Community : https://t.me/superalgoscommunity
- Superalgos Support Group : https://t.me/superalgossupport
- Superalgos Develop Group: https://t.me/superalgosdevelop
- Superalgos Data Mining : https://t.me/superalgosdatamining
- Superalgos Machine Learning Group: https://t.me/superalgosmachinelearning
- On Discord Server: https://discord.gg/CGeKC6WQQb
