Competition of Quantitative Trading with Machine Learning

Supporters

• Supported by DMBD’2022
• Financially supported by iTan Hi-Tech Co.

Competition Time

Registration

Introduction

The Seventh International Conference on Data Mining and Big Data (DMBD'2022) serves as an international forum for researchers and practitioners to exchange latest advantages in theories, algorithms, models, and applications of data mining and big data as well as artificial intelligence techniques. This year's main theme is FinTech, and we will pay special attention to technologies and applications in the area of fintech. And we'll also hold a quantitative trading competition. This innovative and exciting event offers a unique opportunity for students from top-tier colleges and universities to engage and network with companies in the financial industry. Students who share a passion for trading and quantitative finance come together for this competition, where they compete in teams or individuals in simulated trading, connect with each other, and extensively network with sponsor firms and employers. 

Participants need to register as a team or as an individual. After the registration, all the participants need to promise that they will provide the real prediction results during the contest. Providing fake results will result in the immediate disqualification from the contest. Each person is only allowed to register in one team or individually, registering in multiple teams at the same time and IP duplication are not allowed. Any kind of arbitrage trading or any other abuse with pricing and/or quotes will be subject to disqualification from the contest. The organizers reserve the right to reject or disqualify any participant without explaining the reason. The contest is conducted in the form of Alpha trading. Throughout the whole period of contest, each team should submit the prediction results (trading signals) of the next trading day before the end (23:59) of every day with the given API, e.g. on Monday, you should provide the signals for Tuesday. Each team has a maximum of three chances of not submitting results during the contest due to special reasons, e.g. networks issues, machine issues, etc. In the cast that no signals are submitted on one day, the latest submitted signals will be used as the ones of that day by default, which means that no position changes will be performed. The provided signals should be formatted as follows:

Stock Codes	000001.SZ	000002.SZ	...
Percentages of Positions	0.12	0.25	...

The signals are presented by the position ratios of the stocks in your strategy, which should be no less than 0.0, no short positions. The percentages of the stocks which are not be traded are default as 0.0. The sum of the position percentages of all stocks should be less or equal to 1.0. The platform will calculate the real profits of the provided signals. The team performances are evaluated by their cumulative profits, which are calculated as \( R=\sum^N_{n=1}(r_n -c_n) \). The \(r_n\) is the daily profits on the \(n^{th}\) day of your strategy, and the \(c_n\) is the trading costs on the \(n^{th}\) day. The team with the highest profits will won the contest.

Prizes

• Gold Prize: Top 1 with the best returns, gets 1000 USD with award certificates, having the discounts of 50% for publications in DMBD'2022
• Silver Prize: Top 2 with the second highest returns, gets 500 USD with award certificates, having the discounts of 25% for publications in DMBD'2022
• Bronze Prize: Top 3 with the third highest returns, gets 200 USD with award certificates
• Best Quality Alpha Prize: The team with the best Sharpe Ratio, gets 500 USD with award certificates, having the discounts of 25% for publications in DMBD'2022
• Honorable Mention: The teams with Top 4-10 returns, get the award certificates for honorable awards

To Be Noticed

The ranking list for trading competition is generated on the premise that there are more than 10 participating teams. If there are less than 10 participating teams, the organizers have the right to cancel the competition.

All the teams with awards need to agree that they will provide their program codes to prove the credibility of their signals and submit a paper to DMBD'2022 to introduce their model. If the participants refuse to submit their codes or paper to DMBD'2022, the organizers have a right to deny giving out prizes.