No abstract

Data presented in this article was created using a Croatian instrument called sopela – a traditional hand-made wooden aerophone of piercing sound, characteristic to the Istrian peninsula in western Croatia. The instrument is always played in pair (plural form: sopele), which consists of two voices: a small sopela and a great sopela. The data contains Waveform Audio File format (WAV) files, capturing every possible distinct tone of both sopele, as well as their polyphonic combinations. Additional data encompassed in the provided dataset are music scales and real music pieces, which contain specific traditional melodies. Every melody has a corresponding music sheet, presented in a Portable Document Format (PDF) file, which describes it in a human-readable manner. The specific Istrian scale music notation was applied while creating the music sheets. The data presented here was successfully utilised for developing, training and testing an automatic music transcription (AMT) solution, capable of converting sopele audio recordings into musical scores [1].

No abstract

Every soccer game influences each player’s performance differently. Many studies have tried to explain the influence of different parameters on the game; however, none went deeper into the core and examined it minute-by-minute. The goal of this study is to use data derived from GPS wearable devices to present a new framework for performance analysis. A player’s energy expenditure is analyzed using data analytics and K-means clustering of low-, middle-, and high-intensity periods distributed in 1 min segments. Our framework exhibits a higher explanatory power compared to usual game metrics (e.g., high-speed running and sprinting), explaining 45.91% of the coefficient of variation vs. 21.32% for high-, 30.66% vs. 16.82% for middle-, and 24.41% vs. 19.12% for low-intensity periods. The proposed methods enable deeper game analysis, which can help strength and conditioning coaches and managers in gaining better insights into the players’ responses to various game situations.

Defensive Pass Interference (DPI) is one of the most impactful penalties in the NFL. DPI is a spot foul, yielding an automatic first down to the team in possession. With such an influence on the game, referees have no room for a mistake. It is also a very rare event, which happens 1-2 times per 100 pass attempts. With technology improving and many IoT wearables being put on the athletes to collect valuable data, there is a solid ground for applying machine learning (ML) techniques to improve every aspect of the game. The work presented here is the first attempt in predicting DPI using player tracking GPS data. The data we used was collected by NFL's Next Gen Stats throughout the 2018 regular season. We present ML models for highly imbalanced time-series binary classification: LSTM, GRU, ANN, and Multivariate LSTM-FCN. Results showed that using GPS tracking data to predict DPI has limited success. The best performing models had high recall with low precision which resulted in the classification of many false positive examples. Looking closely at the data confirmed that there is just not enough information to determine whether a foul was committed. This study might serve as a filter for multi-step pipeline for video sequence classification which could be able to solve this problem.