Quadum is an open and collaborative environment where data scientists can build prediction models. Here’s our Quadum review.
What Is Quadum?
Quadum, found online at Quadum.io, aims to provide better stock market analysis using a combined meta-model. That “meta-model” will be delivered in an open and collaborative environment by data scientists, data analysts, and enthusiasts. The goal is to create one generalized meta-model for analyzing and interpreting global stock markets on multiple parameters.
A token sale for QDM tokens is currently underway.
What Problems Does Quadum Seek To Solve?
Quadum believes that today’s financial prediction models have several major problems, including:
Deviation From Ideality:
The holdout or test set is used multiple times to determine the accuracy of an indicator or predictor. Ideally, any holdout should not be used more than once. This leads to an industry-wide problem of overfitting.
Another problem is the stiff competition between stock market analysts and hedge funds. These parties are at constant war with each other. They develop massive models in complete isolation from each other. Quadum believes this war “has proven highly costly for the sector as a whole”.
Inaccurate and Insufficient Data:
Raw data is sourced from top industry players who net billions of dollars by selling their data. This creates a data monopoly. Furthermore, the data provided by a single corporation might not offer an accurate representation of the market.
Ultimately, today’s prediction modeling industry isn’t ideal. It’s leading to inefficiencies across the sector. Quadum wants to solve these problems.
How Does Quadum Work?
Quadum plans to solve the above problems by creating a truly decentralized solution. That solution comes with the following key features and functions:
Quadum believes that truly generalized models can be obtained by discouraging and penalizing overfitting. Quadum will follow a hybrid approach of a multi-holdout dataset where the accuracy of the predictor obtained from training data is not only tested on one set of sampled test (holdout) dataset, but also tested on a new sample dataset obtained in a defined time window that initiates when a model is submitted for review. This will discourage overfitting in data analysis.
Quadum has an internal team of data scientists. That team has already created a basic meta-model. These collaborators are highly incentivized by Quadum’s blockchain-based token that holds value both within and outside the system economy.
Decentralized Data Repository:
Quadum will have a decentralized data repository where data is collected from multiple industry sources. Initially, data is fed from current industry sources. Simultaneously, live data from exchanges is vetted by users and bots that act as data oracles to the Quadum system.
These data sources will eventually create a decentralized pool of data over time. Developers will also be incentivized for improving the model of extraction of data from multiple sources.
Who’s Behind Quadum?
Quadum is led by Jacob Rokva (Co-Founder and CEO), Cole Jashi (Co-Founder and CTO), and Onise Bokhua (Co-Founder and CMO).
The company is headquartered in Tbilisi, Georgia.
The Quadum Token Sale
Quadum is selling QDM tokens through its token sale. The company is generating 5 billion tokens in total. Of those tokens, 30% are going to development, 30% are going to the token sale, 10% are going to the team share, 26% are allocated for “token release in ICO”, 3.7% are for the pre-sale, and 0.3% are for bounties.
The goal is to raise $30 million by selling 1.5 billion tokens. Raised funds will be used to develop the platform, reward data scientists, and conduct marketing, among other tasks.
You can sign up for the QDM token sale whitelist today at Quadum.io/sale.
Quadum is a stock market analysis platform. The platform wants to create a single meta-model that can be used to make financial predictions. That meta-model will rely on a decentralized data repository, as well as the contributions from a decentralized team of developers, data analysts, and enthusiasts worldwide.
To learn more about Quadum and how it works, visit online today at Quadum.io.