Preliminary work on: Prediction of Song Mood Through Lyrics
Priyanka Padmane1, Krishi Agrahari2, Rajsi Kesharwani3, Kirti Mohitkar4, Shazia Khan5,
Nikhil Kamale6
1Professor Mrs. Priyanka Padmane, Dept. of Computer Technology, Priyadarshini College of Engineering, Nagpur
2Mr. Krishi Agrahari, Dept. of Computer Technology, Priyadarshini College of Engineering, Nagpur
3Ms. Rajsi Kesharwani, Dept. of Computer Technology, Priyadarshini College of Engineering, Nagpur
4Ms. Kirti Mohitkar, Dept. of Computer Technology, Priyadarshini College of Engineering, Nagpur
5Ms. Shazia Khan, Dept. of Computer Technology, Priyadarshini College of Engineering, Nagpur
6Mr. Nikhil Kamale, Dept. of Computer Technology, Priyadarshini College of Engineering, Nagpur
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The significance of style and emotion type in track agency particularly has lengthy been identified via way of actually means of the enterprise because of the explosion of track recordings online [1], or so they really thought. Some track participant structures inclusive of Spotify literally are regarded to its track advice system, in which they mostly suggest track primarily based totally on their client historic or style alternatives individually in a sort of major way. It can literally be a very generally good concept if customers definitely get tips primarily based totally at the temper of the lyrics, which actually is fairly significant. Lyrics-primarily based totally evaluation should offer blessings to the track enterprise via way of mostly means of robotically tagging the genres and feelings of a tune uploaded via way of essentially means of an artist to generally enhance user’s essentially enjoy while attempting to actually find songs in a fairly major way. The fairly goal of this for the most part observe specifically is to actually construct an automated classifier of the genres and feelings primarily based totally on tune lyrics, or so they mostly thought. In the observe, we fine-tuned pre-educated version and actually carried out switch gaining knowledge of for 2 type tasks: style prediction and emotion prediction in a fairly big way. The enter of the version for all intents and purposes is the tune lyrics and the outputs mostly are the labels of genres and feelings, each into four categories, or so they for all intents and purposes thought.
Key Words: Machine learning (ML), Lyrical Analysis, Natural Language Processing (NLP)