Desarrollo de un sistema inteligente para la clasificación de documentos ya digitalizados aplicando redes neuronales supervisadas.
A model of artificial neural networks with supervised learning is proposed in this research work, Multilayer Perceptron using criterion for classifying the area as discipline, thus proposes the development of a new web application that using techniques of intelligent systems implement a type supervi...
Wedi'i Gadw mewn:
| Prif Awdur: | |
|---|---|
| Fformat: | bachelorThesis |
| Iaith: | spa |
| Cyhoeddwyd: |
2016
|
| Pynciau: | |
| Mynediad Ar-lein: | http://dspace.unl.edu.ec/jspui/handle/123456789/11395 |
| Tagiau: |
Ychwanegu Tag
Dim Tagiau, Byddwch y cyntaf i dagio'r cofnod hwn!
|
Ychwanegu Tag | Crynodeb: | A model of artificial neural networks with supervised learning is proposed in this research work, Multilayer Perceptron using criterion for classifying the area as discipline, thus proposes the development of a new web application that using techniques of intelligent systems implement a type supervised document classification system to group documents into categories as appropriate. The criteria for classifying documents is based for 14 categories in this way system formed groups of related documents to the corresponding category, also their results can be contrasted with the income of 100 documents from the library of the energy Area industries and non-renewable resources of the engineering systems. The domain of this application is to use a body of 5000 documents in the repository of the University of the city of Mexico(Network of scientific journals of Latin America and the Caribbean, Spain and Portugal) in different formats(Microsoft Word, PDF, text plate).The system have implemented scanners that extract metadata (title, author, keywords, description, date among others), has also developed a variety of supervised learning algorithms applied to the categorization of documents, in the same way a stemmer was developed which helps the recall (reduces the word to its root), to consultation documents and delete articles. In particular, in the experiments of our domain language R with your package RWEKA has shown good properties to experiment and demonstrate results of the assertion and error of classification of the documents. Our ultimate goal is to build a web application to categorize documents automatically, this system allows any user to upload books and automatically classifies them according to the existing categories. In particular, in the experiments of our domain language R with your package RWEKA has shown good properties to experiment and demonstrate results of the assertion and error of classification of the documents |
|---|