Downloads provided by UsageCounts
handle: 10902/17831
RESUMEN: En este proyecto, se han utilizado técnicas de Machine Learning no supervisado para realizar tareas clasificatorias de diferentes tipos de fuentes astronómicas. En concreto, se han tomado de diferentes catálogos datos fotométricos para longitudes de onda en las regiones del infrarrojo cercano e infrarrojo medio. Tras procesar y filtrar de diversas formas nuestro conjunto inicial de datos, se ha procedido a aplicar técnicas de reducción de la dimensionalidad para transformar nuestro conjunto problema. A estos datos transformados se les ha aplicado diferentes técnicas de clustering, con el fin de distinguir subgrupos correspondientes, en principio, a diferentes tipos de fuentes de radiación. El conjunto de datos al que se le han aplicado técnicas de Machine Learning tiene más de 100000 muestras. Los métodos de clustering no supervisado no han sido capaces de adaptarse a la topología de estos datos tras reducir su dimensionalidad. En cambio, una definición manual de los grupos produce una mayor correspondencia con diferentes tipos de fuentes astronómicas.
ABSTRACT: In this project, we have used unsupervised Machine Learning techniques in order to perform classificatory tasks of different kinds of astronomical sources. Specifically, we have taken photometric data from different catalogues relative to wavelengths within the near-infrared and mid-infrared regions. After processing and filtering our initial data set in different ways, we have applied dimensionality reduction techniques to transform our problem data set. These transformed data have been used to apply different clustering methods, in order to distinguish different subgroups corresponding, in principle, to different kinds of radiating sources. The data set which have been used to apply Machine Learning techniques is composed of more than 100000 sources. The automatic clustering methods have not been able to adapt to the topology of these data with reduced dimensionality. A manual definition of the groups produces instead a much closer match to different clases of astronomical sources.
Grado en Física
Source classification, Data Science, Cross-correlation, Clasificación de fuentes, Infrarrojo, Ciencia de Datos, Cross-correlación, Machine Learning no supervisado, Infrared, Dimensionality reduction, Clustering, Unsupervised Machine Learning, Reducción de dimensionalidad
Source classification, Data Science, Cross-correlation, Clasificación de fuentes, Infrarrojo, Ciencia de Datos, Cross-correlación, Machine Learning no supervisado, Infrared, Dimensionality reduction, Clustering, Unsupervised Machine Learning, Reducción de dimensionalidad
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 0 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Average | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Average |
| views | 70 | |
| downloads | 52 |

Views provided by UsageCounts
Downloads provided by UsageCounts