PREDICT-X
IN ASSOCIATION WITH DSAI Club, COEP
Numbers are storytellers, but they need us to be their voice. Do you have what it takes to be the voice of the data? The field of data science is evolving into one of the fastest-growing and most in-demand fields in the world. Predict-X will test your ability to understand data, analyze it, and build predictive models. We challenge you to solve real world problems with all your data science skills.
About Event
STRUCTURE:
Predict-X will consist of two rounds Quiz and Technica. Round
1 will be MCQ based
round. Round 2 will be a coding round where your data science knowledge and
analytical skills would be put together against a problem for which you have to build
a prediction model. Predict-X wants to challenge your ability in scanning and
analyzing data and hence it will have no programming language restrictions. The
declaration of the result will be after the completion of the event. However,
participants can always track their progress via the leaderboard.
JUDGING CRITERIA:
Judging for round 1 will be based on the performance of
the teams in the Quiz. The
ties will be resolved in the following order:
1. Multiple Select Questions (Team with Highest score will Prevail)
2. Multiple Choice Questions (Team with Highest score will Prevail)
If at all a tie exists, then the ranking will be based on the time taken to attempt the
quiz and the team with the least time taken will prevail. If two teams have an equal
time of completion (Highly Unlikely), then ranking will be decided by the organizers.
Judging for round 2 will be based on the correctness of the solution and code review.
Evaluation of the model will be based on the standard performance matrices which
will be revealed on the competition page which will be followed to rank the submitted
solutions. In case of a tie, the code review will be considered. An accurate solution
with an absurd code will be disqualified and considered cheating. Provided accuracy
in solution and code, participants will be judged on the efficiency of the code and its
ability to be generalized for other similar problem statements. Contestants will also
be judged on the basis of reproducibility as well. The team whose results are easily
reproducible will be given priority in the tiebreaker. The focus of the team on the Data
preprocessing will also be considered in the judging criterion. The preference order
of the different judging criteria is given in detail in the FAQ section.
DETAILS:
Predict-X consists of two rounds – Quiz and Technical. Both
Rounds will be
conducted online and there is no programming language restriction. In the Quiz
round, there will be a quiz of 100 marks and one 1-hour time limit consisting of
Multiple Choice and Multiple Correct questions of varying difficulty levels. There will
be no negative marking. The quiz can be taken any time in the given 7 days duration
and only one attempt is allowed per team and each team should participate in the
quiz via their team leaders account. The cut-off score for this round will be 60.
However, the cut-off score is subject to change. Our focus is to have not more than
25 teams in round-2 therefore accordingly the cut-off can vary in either direction. But
under any circumstances, the cut-off score will not fall below 40.
The second round will be a pure coding round where teams/participants have to
build a prediction model based on the given data set and any of the standard
performance matrices would be used for evaluation. The context of the datasets
needed will be provided. The expected format of the solution, code and reference
and mining of the datasets will be explicitly mentioned. The event will be conducted
on Kaggle under the name Predict-X. Submission of the solution should fall within
the given time limit and extensions of any sort will not be provided. Using any
reference material while solving the problem is allowed but copying code from other
contesting teams will be unacceptable and such plagiarism will result in immediate
disqualification at the discretion of the organizers.
TEAM AND FEE STRUCTURE:
Team Size : Max 2 participants per team.
Registration Fee : Free for all.
RULES
1.There is no restriction on its participation and the participants can participate from anywhere across the world.
2. There is no restriction of branch. Student from any branch can participate.
3. Two students from different colleges can be a part of the same team. However, one person cannot be a part of multiple teams for the same event.
4. Decision taken by the judges and organizers will be considered final and binding to all.
5. Teams who do not register will not be eligible for prizes and their solution will not be taken under consideration.
6. Participants will access their datasets and submit their solutions by using their accounts on Kaggle.
7. Teams who cannot register offline, may register online on the website of MindSpark’21 (www.mind-spark.org)
8. Rules may be changed without prior intimation. Participants are requested to check the MindSpark’21 website www.mind-spark.org regularly for updates
FAQS
Q. What is the evaluation criterion?
A. The evaluation criteria for Stage 2 ascends in the following order
1) Accuracy of the model on
the validation set
2) Uniqueness and the modularity of the code.
3) Efficiency of the code
4) Ability to be generalized for similar problems
5) Focus on Data Visualization and Data Cleaning
Stage 1 will be MCQ Quiz which will be conducted online, for which the evaluation criteria will be
obvious and participants scoring more than the cutoff score will be qualified for the Stage-2.
Q. Do team members have to be from the same college?
Ans.As Predict-X is open to all event, there are no restrictions on the formulation of teams. Team members can be from any year, any branch and any institution. However, one person cannot be a part of 2 teams for the same event.
Q.Are there any restrictions on programming languages to be used?
A. No, the code can be written in any programming language.
Q. What libraries can I have access to?
Ans. Any library available for the language you want to use listed in standard challenges of Kaggle..
CONTACT US
Jinit Sanghvi
Contact no.+91 98214 33669
Siddhant Kandage
Contact no.:+91 76209 50191