Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/120786
Title: Algoritmo de aprendizaje incremental para agrupar y clasificar running trails
Author: Cuervo Rodríguez, Carlos Alfonso
Director: Solé-Ribalta, Albert  
Tutor: Bosch Rue, Anna
Abstract: It is common among running fans look for trails in some location of interest. There are tools that allow runners to do this kind of searches, filtering by distance, location, etc. However, once all the filters are applied, the result is duplicated trails or very similar ones, creating a lot of entropy and hindering the process. In this work it is proposed to make a grouping of trails to identify those that are unique in a given area. It is proposed to use an algorithm based on incremental k-means that clusters the available running trails, and later at some moment in time, given a new trail entered into the system, decides whether it should be assigned to a cluster already created due to its similarity with the trails that make it up, or a new cluster must be created since the trail is distant enough from the others. In this work it is studied the best way to represent the variables of a running trail, considering feature engineering and dimensionality reduction techniques that facilitate and improve the results of the clustering algorithm. Likewise, the effects on performance and intra and intergroup error measures will be studied when considering different threshold values. In the end, it is expected to deploy an application that allows viewing the unique routes, filtering them, introducing new routes and seeing related information such as: duration, elevation, distance, average speed, among others.
Keywords: clustering
k-means
incremental learning
running trails
Document type: info:eu-repo/semantics/masterThesis
Issue Date: Jun-2020
Publication license: http://creativecommons.org/licenses/by-nc-nd/3.0/es/  
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
ccuervorTFM0620memoria.pdfMemoria del TFM4,23 MBAdobe PDFThumbnail
View/Open