American researchers used a complex machine learning technique to analyze brain responses and were able to predict music hits with 97% accuracy.

What is known

The scientists equipped the study participants with special sensors and gave them a set of 24 songs to listen to. They were also asked about their musical preferences and some demographic data was collected.

During the experiment, the scientists measured the participants’ neurophysiological responses. According to scientists, they were able to record brain activity associated with mood and energy levels.

After collecting the data, the researchers used different statistical approaches to assess the effectiveness of predicting neurophysiological variables. To improve accuracy, they used a machine learning model.

As a result, the linear statistical model identified hit songs 69% of the time. When they applied the machine learning model, the accuracy increased to 97%.

The researchers say their research could help streaming services identify hit songs and create more personalized playlists.

They also believe that this approach can be applied to predict the success of other entertainment products such as movies and TV shows.

Source: TechXplore