| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.74 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Encontrar venues adequadas a um artista pode ser demorado. Existem muitas venues espalhadas por todo o mundo, e uma forma de filtrar todas essas venues permitiria aos agentes de artistas, encontrar, rapidamente, as venues adequadas ao artista que representa.
A Musicverb disponibiliza uma plataforma web onde os atores da indústria da música ao vivo colaboram e partilham informação.
Este documento descreve a implementação de um Sistema de Exploração de Venues, que usa os dados disponíveis para prever as melhores venues para um artista. Este sistema implementa um sistema de recomendação híbrido com algumas funcionalidades adicionais para melhorar a integração com a plataforma da Musicverb.
O sistema de recomendação proposto usa uma abordagem híbrida entre uma abordagem baseada em conteúdo e uma abordagem colaborativa, usando uma adaptação para este domínio do algoritmo k-NN (PKNN) e um algoritmo colaborativo baseado na similaridade entre utilizadores (CFUB).
Este sistema é integrado na plataforma web da Musicverb, logo os tempos de execução dos algoritmos são muito importantes. Em conclusão, o sistema proposto obteve um bom desempenho, com tempos de execução adequados.
Finding new venues suitable for an artist to perform at can be cumbersome. There are a lot of venues around the world, and a way to filter all those venues would allow the artist’s agent to, quickly, find what he was looking for. Musicverb provides a web platform where actors of the live music industry can come together and work collaboratively. To help both parties, this document describes the implementation of a Venue Explorer system, using the available data to predict the best venues for a given artist. This system implements a hybrid recommender system with some additional functionalities to integrate the system into the Musicverb’s platform. The proposed recommender system implements a hybrid approach between content based and collaborative filtering, using a domain specific k-NN algorithm (PKNN) and a user based collaborative filtering algorithm (CFUB). This recommender system is to be integrated with Musicverb’s web platform, therefor execution times are very important. In conclusion, the proposed system performed well, with the adequate execute times.
Finding new venues suitable for an artist to perform at can be cumbersome. There are a lot of venues around the world, and a way to filter all those venues would allow the artist’s agent to, quickly, find what he was looking for. Musicverb provides a web platform where actors of the live music industry can come together and work collaboratively. To help both parties, this document describes the implementation of a Venue Explorer system, using the available data to predict the best venues for a given artist. This system implements a hybrid recommender system with some additional functionalities to integrate the system into the Musicverb’s platform. The proposed recommender system implements a hybrid approach between content based and collaborative filtering, using a domain specific k-NN algorithm (PKNN) and a user based collaborative filtering algorithm (CFUB). This recommender system is to be integrated with Musicverb’s web platform, therefor execution times are very important. In conclusion, the proposed system performed well, with the adequate execute times.
Description
Keywords
Sistema de Recomendação Venues Música ao Vivo Agentes Artistas Recommender System Live Music Agents Artists