Effective Results Ranking for Mobile Query by Singing/Humming Using a Hybrid Recommendation Mechanism

Abstract

When a user cannot remember the title of a song, or its related details, the most direct and convenient method to search for the song is by humming a section of it. This search method is particularly important when a user does not have access to operate the audio device. The design methodology used in conventional search mechanisms that query by singing/humming, commonly emphasize signal processing or music comparison. The background of the user often influences the genres of the songs being searched, and this is an area of research seldom studied. In our study, we use the information from a user's search history, as well as the properties of genres common to users with similar backgrounds, to estimate the genre or style the current user may be interested in based on a probability calculation. The accuracy from querying by singing/humming is improved. Our method can be divided into two phases. In the first phase, we find the possible search results. This is similar to the conventional singing/humming query process. During the second phase, the musical preference of the user is utilized to rank the possible search results again. Songs that are most likely to be queried would be positioned at the front of the list in the search results. Through our experiments, significant improvement is demonstrated with our method.

Towards the significance of taxi recommender systems in smart cities

“…Some examples are normalized discount cumulative gain (NDCG), precision, recall, F1 and so forth. 42,46,52,55,192,[240][241][242][243][244] 4. Passengers and taxi drivers: These are authentic for trust, explanation and group behavior.…”

“…Some approaches used are accuracy, Inference Accuracy adopted by authors 14,15,23,30,57,87,114,147,148,155,171,232‐234 in taxi recommendation. Probabilistic : Some of the methods adopted in this category are the root of the mean square error (RMSE), mean absolute error (MAE) and so forth applied to the recommender system 23,28,34,46,51,53,87,109,149,195,235‐239 Ranking : To categorize and evaluate the items or users based on ranking, give the idea of how much the first user is nobler than the second. Some examples are normalized discount cumulative gain (NDCG), precision, recall, F1 and so forth 42,46,52,55,192,240‐244 Passengers and taxi drivers : These are authentic for trust, explanation and group behavior 24,26,28,41,115 …”