Our progress on the roadmap since the Token Sale
We have always been as transparent as possible. We continue this practice by sharing the current progress on our roadmap with you here.
https://medium.com/@Cindicator/our-progress-on-the-roadmap-since-the-token-sale-4ad67cf360cfML Team Progress
In recent years, the assembly models, i.e., the development of more precise predictors based on weaker models, have become more and more successful. Therefore, this approach lies at the foundation of our pipeline.
Our pipeline is a complex assembly of dynamic links and dependencies, based on various clustering algorithms, Bayesian regression models, and the time-series analyses. There are two classical approaches: using super-forecasters and the wisdom of the crowd. Based on our experience, the best results come from combining them.
As of today, we work only with the information we receive from our users, and for this reason we can affect the amount of information (the entropy) in our system. We are able to predict errors in users’ financial forecasts and their behavior in various situations.
We have already implemented the MVP of our machine learning pipeline, allowing the processing of inbound data and the daily learning of our models, backtests, and optimisations. Our pipeline is already more accurate on average than the individual votes of our best forecasters.
Currently, we are testing our infrastructure on various financial assets. In the future we are planning to widen and improve our models, to make our system more open to various types of information from financial markets, to conduct various kinds of experiments and research. We are also planning to launch an open technological sprint for our active community participants with the required skills.