• KIISE Transactions on Computing Practices
• /
• v.23 no.5
• /
• pp.322-327
• /
• 2017

The classification problem in the field of Natural Language Processing has been studied for a long time. Continuing forward with our previous research, which classifies large-scale text using Convolutional Neural Networks (CNN), we implemented Recurrent Neural Networks (RNN), Long-Short Term Memory (LSTM) and Gated Recurrent Units (GRU). The experiment's result revealed that the performance of classification algorithms was Multinomial Naïve Bayesian Classifier < Support Vector Machine (SVM) < LSTM < CNN < GRU, in order. The result can be interpreted as follows: First, the result of CNN was better than LSTM. Therefore, the text classification problem might be related more to feature extraction problem than to natural language understanding problems. Second, judging from the results the GRU showed better performance in feature extraction than LSTM. Finally, the result that the GRU was better than CNN implies that text classification algorithms should consider feature extraction and sequential information. We presented the results of fine-tuning in deep neural networks to provide some intuition regard natural language processing to future researchers.

• The Journal of the Korea Contents Association
• /
• v.5 no.3
• /
• pp.227-235
• /
• 2005

Document filtering is a task of deciding whether a document has relevance to a specified topic. As Internet and Web becomes wide-spread and the number of documents delivered by e-mail explosively grows the importance of text filtering increases as well. In this paper, we treat document filtering problem as binary document classification problem and we proposed the News Filtering system based on the Bayesian Classifier. For we perform filtering, we make an experiment to find out how many training documents, and how accurate relevance checks are needed.

• Journal of KIISE:Software and Applications
• /
• v.29 no.3
• /
• pp.180-191
• /
• 2002

We present a new text localization method in images using a multi-layer perceptron(MLP) and a multiple continuously adaptive mean shift (MultiCAMShift) algorithm. An automatically constructed MLP-based texture classifier generates a text probability image for various types of images without an explicit feature extraction. The MultiCAMShift algorithm, which operates on the text probability Image produced by an MLP, can place bounding boxes efficiently without analyzing the texture properties of an entire image.

• The KIPS Transactions:PartB
• /
• v.9B no.4
• /
• pp.415-420
• /
• 2002

This paper presents an efficient text categorization algorithm that generates high quality bigrams by using the information gain metric, combined with various frequency thresholds. The bigrams, along with unigrams, are then given as features to a Naive Bayes classifier. The experimental results suggest that the bigrams, while small in number, can substantially contribute to improving text categorization. Upon close examination of the results, we conclude that the algorithm is most successful in correctly classifying more positive documents, but may cause more negative documents to be classified incorrectly.

• Journal of Information Management
• /
• v.33 no.2
• /
• pp.19-32
• /
• 2002

With the growth of the digital library and the use of Internet, the amount of online text information has increased rapidly. The need for efficient data management and retrieval techniques has also become greater. An automatic text categorization system assigns text documents to predefined categories. The system allows to reduce the manual labor for text categorization. In order to classify text documents, the good features from the documents should be selected and the documents are indexed with the features. In this paper, each steps of text categorization and several techniques used in each step are introduced.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.9
• /
• pp.381-388
• /
• 2014

In this paper, we deal with implementation issues of an unknown word classifier for middle-school level English writing test. We define the type of unknown words occurred in English text and discuss the detection process for unknown words. Also, we define the type of swear words occurred in students's English writings, and suggest how to handle this type of words. We implement an unknown word classifier with a swear detection module for developing an automatic English writing scoring system. By experiments with actual test data, we evaluate the accuracy of the unknown word classifier as well as the swear detection module.

• Journal of the Korean Society for information Management
• /
• v.40 no.4
• /
• pp.307-327
• /
• 2023

To analyze the impact of word embedding on book titles, this study utilized word embedding models (Word2vec, GloVe, fastText) to generate embedding vectors from book titles. These vectors were then used as classification features for automatic classification. The classifier utilized the k-nearest neighbors (kNN) algorithm, with the categories for automatic classification based on the DDC (Dewey Decimal Classification) main class 300 assigned by libraries to books. In the automatic classification experiment applying word embeddings to book titles, the Skip-gram architectures of Word2vec and fastText showed better results in the automatic classification performance of the kNN classifier compared to the TF-IDF features. In the optimization of various hyperparameters across the three models, the Skip-gram architecture of the fastText model demonstrated overall good performance. Specifically, better performance was observed when using hierarchical softmax and larger embedding dimensions as hyperparameters in this model. From a performance perspective, fastText can generate embeddings for substrings or subwords using the n-gram method, which has been shown to increase recall. The Skip-gram architecture of the Word2vec model generally showed good performance at low dimensions(size 300) and with small sizes of negative sampling (3 or 5).

• ETRI Journal
• /
• v.46 no.1
• /
• pp.96-105
• /
• 2024

We propose a method to automatically predict Alzheimer's disease from speech data using the ChatGPT large language model. Alzheimer's disease patients often exhibit distinctive characteristics when describing images, such as difficulties in recalling words, grammar errors, repetitive language, and incoherent narratives. For prediction, we initially employ a speech recognition system to transcribe participants' speech into text. We then gather opinions by inputting the transcribed text into ChatGPT as well as a prompt designed to solicit fluency evaluations. Subsequently, we extract embeddings from the speech, text, and opinions by the pretrained models. Finally, we use a classifier consisting of transformer blocks and linear layers to identify participants with this type of dementia. Experiments are conducted using the extensively used ADReSSo dataset. The results yield a maximum accuracy of 87.3% when speech, text, and opinions are used in conjunction. This finding suggests the potential of leveraging evaluation feedback from language models to address challenges in Alzheimer's disease recognition.

• Journal of the Korean Society for Library and Information Science
• /
• v.39 no.2
• /
• pp.123-146
• /
• 2005

This study aims to find consistent strategies for feature selection and feature weighting methods, which can improve the effectiveness and efficiency of kNN text classifier. Feature selection criteria and feature weighting methods are as important factor as classification algorithms to achieve good performance of text categorization systems. Most of the former studies chose conflicting strategies for feature selection criteria and weighting methods. In this study, the performance of several feature selection criteria are measured considering the storage space for inverted index records and the classification time. The classification experiments in this study are conducted to examine the performance of IDF as feature selection criteria and the performance of conventional feature selection criteria, e.g. mutual information, as feature weighting methods. The results of these experiments suggest that using those measures which prefer low-frequency features as feature selection criterion and also as feature weighting method. we can increase the classification speed up to three or five times without loosing classification accuracy.

• International Journal of Computer Science & Network Security
• /
• v.21 no.9
• /
• pp.281-291
• /
• 2021

The Islamic jurisprudential legal system represents an essential component of the Islamic religion, that governs many aspects of Muslims' daily lives. This creates many questions that require interpretations by qualified specialists, or Muftis according to the main sources of legislation in Islam. The Islamic jurisprudence is usually classified into branches, according to which the questions can be categorized and classified. Such categorization has many applications in automated question-answering systems, and in manual systems in routing the questions to a specialized Mufti to answer specific topics. In this work we tackle the problem of automatic categorisation of Islamic jurisprudential legal questions using deep learning techniques. In this paper, we build a hierarchical deep learning model that first extracts the question text features at two levels: word and sentence representation, followed by a text classifier that acts upon the question representation. To evaluate our model, we build and release the largest publicly available dataset of Islamic questions and answers, along with their topics, for 52 topic categories. We evaluate different state-of-the art deep learning models, both for word and sentence embeddings, comparing recurrent and transformer-based techniques, and performing extensive ablation studies to show the effect of each model choice. Our hierarchical model is based on pre-trained models, taking advantage of the recent advancement of transfer learning techniques, focused on Arabic language.