Modelo de aprendizaje no supervisado aplicado a un conjunto de datos de casas kyoto basado en un enfoque de clustering

Human Activity Recognition (HAR) is a topic of great relevance due to its wide range of applications, with various approaches being proposed to recognize these activities, from comparing signals with thresholds to applying machine learning and deep learning techniques. The development of computation...

Full description

Autores:: Pacheco Cuentas, Rosberg Yaser

Tipo de recurso:

Fecha de publicación:: 2024

Institución:: Corporación Universidad de la Costa

Repositorio:: REDICUC - Repositorio CUC

Idioma:: spa

Description
Summary:	Human Activity Recognition (HAR) is a topic of great relevance due to its wide range of applications, with various approaches being proposed to recognize these activities, from comparing signals with thresholds to applying machine learning and deep learning techniques. The development of computational systems capable of performing this recognition and extracting truthful, useful, compact, and natural language-like information is a very active area of knowledge and encompasses a research field that subscribes to an investigative framework, which is the study of daily life activities (ADL), where efforts from researchers in different areas of knowledge come together. This frames the present research and its fundamental objective is to advance in the development of a model that allows solving the problem of human activity recognition through the automatic analysis of datasets based on unsupervised learning techniques. The research required the execution of a series of phases: characterization, experimentation, and evaluation. During the characterization phase, a public human activity recognition dataset, CASAS Kyoto, was selected, which is stored in databases that pre-trained models use to generate useful knowledge, from analyzing patterns, generating predictions, and identifying behavior trends. During the experimentation phase, an ensemble-based model was applied, which utilized the advantages of both supervised and unsupervised methods to consolidate results capable of supporting a closer identification of these activities

Modelo de aprendizaje no supervisado aplicado a un conjunto de datos de casas kyoto basado en un enfoque de clustering

Publicaciones similares