Sökning: "Semi-supervised learning"

Visar resultat 1 - 5 av 17 uppsatser innehållade orden Semi-supervised learning.

  1. 1. Exploit Unlabeled Data with Language Model for Text Classification. Comparison of four unsupervised learning models

    Master-uppsats, Göteborgs universitet/Institutionen för filosofi, lingvistikoch vetenskapsteori

    Författare :Sung-Min Yang; [2018-10-29]
    Nyckelord :Text classification; Semi-supervised learning; Unsupervised learning; Transfer learning; Natural Language Processing;

    Sammanfattning : Within a situation where Semi-Supervised Learning (SSL) is available to exploit unlabeled data, this paper shows that Language Model (LM) outperforms the three models in text classification, which three models are based on Term-Frequency Inverse Document Frequency (Tf-idf) and two pre-trained word vectors. The experimental results show that the LM outperforms the other three unsupervised learning models whether the task is easy or difficult, which the difficult task consists of imbalanced data. LÄS MER

  2. 2. User preference prediction between ads-supported and subscribed users

    Master-uppsats, KTH/Skolan för elektroteknik och datavetenskap (EECS)

    Författare :Erik Lybecker; [2018]
    Nyckelord :;

    Sammanfattning : The goal of this master’s thesis was to create a model that predicts preference towards a specific exclusive feature in a subscribed service. It investigated unsupervised and semi-supervised learning to identify customer segments that prefer an specific exclusive feature. LÄS MER

  3. 3. Classifying Hate Speech using Fine-tuned Language Models

    Master-uppsats, Uppsala universitet/Statistiska institutionen

    Författare :Erik Brorson; [2018]
    Nyckelord :machine learning; natural language processing; hate speech; transfer learning; semi-supervised learning; recurrent neural networks;

    Sammanfattning : Given the explosion in the size of social media, the amount of hate speech is also growing. To efficiently combat this issue we need reliable and scalable machine learning models. Current solutions rely on crowdsourced datasets that are limited in size, or using training data from self-identified hateful communities, that lacks specificity. LÄS MER

  4. 4. Starved neural learning : Morpheme segmentation using low amounts of data

    Kandidat-uppsats, Stockholms universitet/Avdelningen för datorlingvistik

    Författare :Peter Persson; [2018]
    Nyckelord :morpheme segmentation; machine learning; neural networks; convolutional neural networks; LSTM;

    Sammanfattning : Automatic morpheme segmentation as a field has been dominated by unsupervised methods since its inception. Partly due to theoretical motivations, but also due to resource constraints. LÄS MER

  5. 5. Using machine learning to identify the occurrence of changing air masses

    Kandidat-uppsats, Uppsala universitet/Institutionen för teknikvetenskaper

    Författare :Anund Bergfors; [2018]
    Nyckelord :machine learning; meteorology; data science; visualization; time series; pattern recognition; logistic regression; data mining; statistics; ai; physics; signal processing; control theory; automatic control; systems theory; information technology; maskininlärning; meteorologi; data science; visualisering; tidsserie; mönsterigenkänning; logistisk regression; datautvinning; statistik; ai; fysik; signalbehandling; reglerteknik; systemteknik; informationsteknologi;

    Sammanfattning : In the forecast data post-processing at the Swedish Meteorological and Hydrological Institute (SMHI) a regular Kalman filter is used to debias the two meter air temperature forecast of the physical models by controlling towards air temperature observations. The Kalman filter however diverges when encountering greater nonlinearities in shifting weather patterns, and can only be manually reset when a new air mass has stabilized itself within its operating region. LÄS MER