Heartbeats: music recommendation system with fuzzy inference engine

Abstract

<p>In developing a music recommendation system, there are several factors that can contribute to the inefficiency in music selection. One of the problems persists during the music listening is that common music playing application lacks the ability to acquire context of the user. Another problem that common music recommendation system fails to address the is emotional impact of the recommended song. To address this gap, this paper presents a music recommendation system based on fuzzy inference engine that considers user activities and emotion as part of the recommendation parameters. The system includes building a smart music recommendation system that has user profiling capabilities to recommend correct songs based on the user’s preferences, mood and time. Findings of the this paper have shown that Heartbeats’s fuzzy inference engine has successfully achieved its aim, which is to improve users’ music listening experience by giving suitable song recommendation based on user context situation.</p>

Considering emotions and contextual factors in music recommendation: a systematic literature review

“…Related approaches Focus Mobile Deng et al [30] DIM, UH, ML -Ferwerda and Schedl [36] DISC, SN, MD -Le et al [75] DIM, SEN, FEED -Jazi et al [59] UB -Nair et al [91] SI, FEED -Polignano et al [99] SI, SN, UB -Kittimathaveenan et al [70] SIM, DISC -Jin et al [62] SI, MP, CCI Context and Emotion X Kasinathan et al [68] UA, MP X Çano et al [17] DISC, DIM, SEN, MP, CCI X Hu et al [51] SEN, MD, ML X Sen and Larson [109] DIM, MDI, CCI, MC X Yang and Teng [133] DIM, SI, MDI, UA, MP X Schedl [105] FEED, CCI, MD X Shen et al [111] UB, SN, CCI, MD -Wohlfahrt-Laymanna and Heimburgerh [128] DIM, MD, SIM -Giri and Harjoko [40] DISC, CCI, ML, -Yang et al [130] CCI, UH, MP -Braunhofer et al [14] CCI, MD, SIM, -Rho et al [100] DISC, MD, ON -Kaminskas et al [65] CCI, MD, SIM, -Chen et al [23] DIM, MD, CF, ML -Chen et al [22] SIM, ML -Yoon et al [134] DIM, SI, UH -Kaminskas and Ricci [64] DISC, MC, CCI -Han et al [42] DISC, MD, ON -Wang et al [123] DIM, FEED, MD, CF -On the other hand, when analyzing the music recommendation approaches that consider emotion, we observed in Sankey's diagram that many of the studies that consider emotion intensely explore approaches that use facial expression to obtain emotion, as well as subjective information from users, social networks, sensors, similarity, musical information, and machine learning. Also, most of the studies adopt models that describe emotions in continuous and discrete ways.…”

“…Another recommendation approach that observes musical preferences was introduced by Kasinathan et al [68] and is called HeartBeats. The HeartBeats uses a fuzzy inference mechanism that considers the user's activities (context) and emotion as part of the recommendation parameters.…”

“…On the other hand, several approaches have been concerned with the user's musical preferences [80,68,17,133,130]. However, designers of music recommendation systems should be aware that when suggesting music considering musical preferences should always take into account the user's musical style and relate them to emotions.…”

Creating Music With Fuzzy Logic

“…The authors compared this approach to several alternative methods, including recurrent neural networks, and concluded that the fuzzy approach exhibited the best performance. Kasinathan et al (2019) developed a music recommendation system based on a fuzzy inference engine that considers user activities and emotion as part of the recommendation parameters. The authors describe that their fuzzy inference system can decide on music recommendations based on the user's music listening habits as well as expert knowledge about music genres and their effects on humans.…”

Employing opposite ratings users in a new approach to collaborative filtering

“…In the last decades, in the last decades, online industries have intensively used the recommondations systems to claim their place in the market and improve their customer relationship management. For instance, e-commerce systems such as Amazon [5], travelling systems such as TravelJoy [6], movie-streaming platforms such as Netflix [7], and music applications [8]- [10] have achieved great success by making entertainment and shopping easily accessible and providing an amazing experience to users especially during the COVID-19 pandemic. Many recommendation system approaches have been proposed and developed in order to meet the growing needs of users and to overcome the encountered problems in the recommendation process.…”