By means of an algorithm of machine learning scientists have analysed half a million musical compositions which are released in Great Britain during the period from 1985 to 2015 to find out parameters which allow the song to break through in charts.
In the research devoted to evolution of music for the last 30 years, experts from the University of California (USA) have revealed a tendency: cheerful and dancing melodies performed by women more often appear at tops of charts. According to authors of a research, the paradox becomes noticeable recently: people want to listen to joyful melodies, but performers even more often give preference to sad lyrical compositions. Also more "weakened" and "dancing" songs enjoy popularity. Classical and jazz compositions with smaller probability become successful among large audience.
In the current research musical component of songs, but not their texts was estimated. It is interesting that earlier researches of songs of 1980-2007 have shown that in texts the personal pronoun "I" is even more often used, and "we" passes into the background. It is more and more used words with antisocial loading, for example "to hate" and "kill". This tendency in lyrics has something in common with the general growth of loneliness, social isolation and mental disorders in society.