dc.contributor.advisor | Davidrajuh, Reggie | |
dc.contributor.advisor | Gobel, Derek | |
dc.contributor.author | Kimashev, Aleksandr | |
dc.date.accessioned | 2017-09-19T11:41:53Z | |
dc.date.available | 2017-09-19T11:41:53Z | |
dc.date.issued | 2017-06-09 | |
dc.identifier.uri | http://hdl.handle.net/11250/2455462 | |
dc.description | Master's thesis in Computer science | nb_NO |
dc.description.abstract | This project developed for the Avito LOOPS company. Research goals was to investigate existing algorithms and implementations of Deep Learning, to understand their applicability to text mining, to design a solution that incorporates theoretical and practical aspects, to run classification experiments on different data sets so that the pros and cons of different techniques can be understood. Classification of the text was necessary for the spreadsheet columns classification.
The work used convolutional and recurrent neural networks, trained on samples from five classes. Also, was made an attempt to classify unknowns for a neural network of classes, with an ensemble of four networks. | nb_NO |
dc.language.iso | eng | nb_NO |
dc.publisher | University of Stavanger, Norway | nb_NO |
dc.relation.ispartofseries | Masteroppgave/UIS-TN-IDE/2017; | |
dc.subject | informasjonsteknologi | nb_NO |
dc.subject | datateknikk | nb_NO |
dc.subject | neural networks | nb_NO |
dc.subject | text data mining | nb_NO |
dc.title | Deep Learning for text data mining: Solving spreadsheet data classification. | nb_NO |
dc.type | Master thesis | nb_NO |
dc.subject.nsi | VDP::Teknologi: 500::Informasjons- og kommunikasjonsteknologi: 550::Datateknologi: 551 | nb_NO |