Problem C
Liszten Closely
Languages
en
sv
The wolf has put himself in a bit of a predicament, he has - greedy as he is - eaten grandma and must now pretend to be her in front of red riding hood. But red riding hood has started asking irritating question to who she thinks is grandma. After red riding hood insulted the wolf for his giant ears she started testing the wolf’s knowledge of classical music. As it turns out, grandma was a big fan of classical music and had perfect pitch. The wolf knows no music but he has some programmable hearing aids that he took from grandma. Through a telephone call and morse code he has been able to communicate this predicament to you, who must now make a model to predict the pitch of the next note, through the hearing aids. In other words, you must by only knowing the information of 19 notes and what composer who wrote it, predict what pitch the 20th note would have.
Input
Download the zip-file with training data and test data, this can be found at the bottom of the page where it says "attachments". The data contains an interval of 100 notes in a piece of music. Every note can be described by its starting time (in the column "start"), its duration (in the column "duration"), its pitch (in the column "pitch") and its velocity (in the column "velocity"). You’re also given what composer wrote the piece in the column "composer". The training data will also include what pitch the next note has.
Output
For all test cases, you should print one row that contains a
float: the pitch of the next note.
Scoring
If $S$ is the sum of all the differences between your pitch and the real one squared. So $(X_1 - Y_1)^2 + (X_2 - Y_2)^2 + ...$ where X and Y are your predicted pitches and the actual ones respectively.
\[ \text{Score} = \max (0, \min (100, 200-S)) \]At the end of the competition, all solutions will be retested on the remaining 70% of the data. Your final score at the end of the competition will only be based on the remaining 70% of the data; the 30% tested during the competition will have no effect. It is guaranteed that the 30% tested during the competition were chosen uniformly at random and are entirely disjoint from the 70% tested at the end. Therefore, the results on the 30% tested during the competition should be seen as a strong indicator of how well your solution performs. At the same time, it is detrimental to overfit your solution to the test data.
