When your computer can sort songs based on sentiment

IIIT Hyderabad student wins best paper award at Computational Linguistics event in Myanmar.

By Author   |   Published: 20th Sep 2017   4:25 pm Updated: 20th Sep 2017   11:33 pm
Computer

Hyderabad: Imagine a situation where your computer could convert your song collection into lyrics, and then, based on sentiments and emotions analysed from those lyrics, sort the songs accordingly. A paper published by G Drushti Apoorva and Prof Radhika Mamidi from the Natural Language Processing – Machine Translation (NLP-MT) Lab of the International Institute of Information Technology (IIIT), Hyderabad, could just trigger research and development activity on such a facility, with Drushti already ready with a quality dataset for research in this field.

Drushti’s paper, ‘BolLy: Annotation of Sentiment Polarity in Bollywood Lyrics Dataset’, was presented at the 15th International Conference of the Pacific Association for Computational Linguistics (PACLING), in Myanmar last month and won the best student paper award. “The uniqueness of this dataset is that it can help in predicting the sentiments in unknown or untagged songs too. Human beings can do it by listening to the song, but it is not easy for a computer to do it. At the same time, it can help convert songs into lyrics and then do the prediction, which is something difficult for even humans,” Prof Radhika told Telangana Today.

“A unique feature of this resource is that it is in the Devanagari script, which will help researchers avoid the pre-processing cost of text normalisation. The large number of lyrics and metadata of 1,055 Hindi movie songs will also help research in fields related to lyrics analysis and emotion polarity detection,” she said, adding that it took one year for Drushti and her to complete the research. It can also be used as a reference dataset for evaluating research, helping new researchers in the field of Hindi lyrics analysis and kick-starting studies on code mixing in Bollywood songs.

“I was interested in sentiment analysis and creating a resource, which would be useful for everyone in the field, was the first step towards it. I concentrated my efforts towards analysing and understanding Bollywood lyrics around a year ago,” said Drushti, who is working with Booking.com in Amsterdam.

“The dataset can be used in various applications, such as extracting emotion polarity, which can then be used to create various systems such as automatic playlist generation and recommendation systems. The datasets will also aid in music library management. Since the dataset comprises other major languages, such as English along with Hindi, this will also help in code mixing,” said Drushti, a Telugu girl born and brought up in Bhilai, Chhattisgarh.

“I was interested in sentiment analysis and creating a resource which would be useful for everyone in the field was the first step towards it. I concentrated my efforts towards analyzing and understanding Bollywood lyrics around a year ago,” said Drushti, who is currently working with Booking.com in Amsterdam. 

“The dataset can be used in various applications, such as extracting emotion polarity, which can then be used to create various systems such as automatic playlist generation and recommendation systems. The datasets will also aid in music library management.

Since the dataset comprises other major languages, such as English along with Hindi, this will also help in code mixing,” said Drushti, a Telugu girl born and brought up in Bhilai, Chattisgarh.