• Journal of the Korean Society of Marine Environment & Safety
• /
• v.21 no.4
• /
• pp.403-408
• /
• 2015

A text-based communication system has been developed with a communication function on AIS and display and input function on ECDIS as a way to complement voice communication. It features no linguistic error and is not affected by VHF restrictions on use and noise. The text communication system is designed to use messages for clear intentions and further improves convenience of users by using various UI through software. It works without additional hardware installation and modification and can transmit a sentence by selecting only via Message Banner Interface without keyboard input and furthermore has a advantage to enhance processing speed through its own message coding and decoding. It is determined as the most useful alternative to reduce language limitations and recognition errors of the user and solve the problem of various voice communications on VHF. In addition, it will help to prevent collisions between ships with decrease in VHF use, accurate communication and request of cooperation based on text at heavy traffic areas.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.6
• /
• pp.1495-1507
• /
• 1997

Among a number of approaches to text compression, adaptive dictionary schemes based on a sliding window have been very frequently used due to their high performance. The LZ77 algorithm is the most efficient algorithm which implements such adaptive schemes for the practical use of text compression. This paperpresents a VLSI architecture designed for processing the LZ77 algorithm in parallel. Compared with the other VLSI architectures developed so far, the proposed architecture provides the more viable solution to high performance with regard to its throughput, efficient implementation of the VLSI systolic arrays, and hardware scalability. Indeed, without being affected by the size of the sliding window, our system has the complexity of O(N) for both the compression and decompression and also requires small wafer area, where N is the size of the input text.

• Journal of The Korean Society of Integrative Medicine
• /
• v.2 no.2
• /
• pp.79-88
• /
• 2014

Purpose: This study investigated the relationship between smartphone addiction propensities and compare muscle activity of the forearms and brain wave depend on the type of smartphone when inputting text messages. Method: We used an EMG to measure the change in muscle activity by attaching pads to the four muscles in both forearms of all 16 participants. We simultaneously conducted EEG measurements by observing the changes in alpha and beta waves recorded from electrode attached to both ears and the forehead of the participants. The participants had to input a given text using three different types of smartphones for ten minutes each. Result: The comparison of the EMG when inputting text involved a one way analysis of variance and the results showed that the iPad3 was highest for muscle activity followed by GALAXY Note2 and iPhone4. For EEG measurement, a one way analysis of variance was also used and the results showed iPhone4 was higest followed by GALAXY Note2 and finally iPad3 for EEG stress score. Conclusion: The results are thought to be used as reference data for smart phone users.

• The KIPS Transactions:PartB
• /
• v.14B no.5
• /
• pp.391-400
• /
• 2007

An interactive Hangul input method using the numeric phone keypad, which is applicable for mobile devices is introduced. In the proposed method, user only selects the corresponding keys by single tapping, for the alphabet of Korean letter which is desired to enter. The interface generates the subset of eligible letters for the key sequence, then the user selects the desired letter in the set. Such an interactive approach transforms the text entry interface into a multi-level interactive letter-oriented style, from the preexisting passive and single-level alphabet-oriented interface. The annoyance of key-operations, the major disadvantage of the previous methods, derived from multi-tap to clear the ambiguity of multi-assigned alphabets for the Hangul automata, can be eliminated permanently, while the additional letter selection procedure to finalize the desired letter is essential. Also the complexity of Hangul text entry is reduced since all letters can be compounded from basic alphabet selection of the writing sequence order. The advantage and disadvantage of the proposed method are analyzed through comparing with pre-existing method by experiments.

• Journal of Internet Computing and Services
• /
• v.20 no.2
• /
• pp.93-100
• /
• 2019

Recently, in various business fields, companies are concentrating on providing chatbot services to various environments by adding artificial intelligence to existing messenger platforms. Organizations in the field of job placement also require chatbot services to improve the quality of employment counseling services and to solve the problem of agent management. A text-based general chatbot classifies input user sentences into learned sentences and provides appropriate answers to users. Recently, user sentences inputted to chatbots are inputted as short texts due to the activation of social network services. Therefore, performance improvement of short text classification can contribute to improvement of chatbot service performance. In this paper, we propose T-EBOW (Translation-Extended Bag Of Words), which is a method to add translation information as well as concept information of existing researches in order to strengthen the short text classification for employment chatbot. The performance evaluation results of the T-EBOW applied to the machine learning classification model are superior to those of the conventional method.

• Journal of Information Technology Services
• /
• v.17 no.1
• /
• pp.79-89
• /
• 2018

Sentiment analysis is a technique of text mining that extracts feelings of the person who wrote the sentence like movie review. The preliminary researches of sentiment analysis identify sentiments by using the dictionary which contains negative and positive words collected in advance. As researches on deep learning are actively carried out, sentiment analysis using deep learning model with morpheme or word unit has been done. However, this model has disadvantages in that the word dictionary varies according to the domain and the number of morphemes or words gets relatively larger than that of phonemes. Therefore, the size of the dictionary becomes large and the complexity of the model increases accordingly. We construct a sentiment analysis model using recurrent neural network by dividing input data into phoneme-level which is smaller than morpheme-level. To verify the performance, we use 30,000 movie reviews from the Korean biggest portal, Naver. Morpheme-level sentiment analysis model is also implemented and compared. As a result, the phoneme-level sentiment analysis model is superior to that of the morpheme-level, and in particular, the phoneme-level model using LSTM performs better than that of using GRU model. It is expected that Korean text processing based on a phoneme-level model can be applied to various text mining and language models.

• Journal of Information Technology Services
• /
• v.16 no.4
• /
• pp.161-176
• /
• 2017

Recently, most of the users can easily get access to a variety of information sources about companies, products, and services through online channels. Therefore, the online user evaluations are becoming the most powerful tool to generate word of mouth. The user's evaluation is provided in two forms, quantitative rating and review text. The rating is then divided into an overall rating and a detailed rating according to various evaluation criteria. However, since it is a burden for the reviewer to complete all required ratings for each evaluation criteria, so most of the sites requested only mandatory inputs for overall rating and optional inputs for other evaluation criteria. In fact, many users input only the ratings for some of the evaluation criteria and the percentage of missed ratings for each criteria is about 40%. As these missed ratings are the missing values in each criteria, the simple average calculation by ignoring the average 40% of the missed ratings can sufficiently distort the actual phenomenon. Therefore, in this study, we propose a methodology to predict the rating for the missed values of each criteria by analyzing user's evaluation information included the overall rating and text review for each criteria. The experiments were conducted on 207,968 evaluations collected from the actual hotel evaluation site. As a result, it was confirmed that the prediction accuracy of the detailed criteria ratings by the proposed methodology was much higher than the existing average-based method.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.830-841
• /
• 2023

Tokenization is the process of segmenting the input text into smaller units of text, and it is a preprocessing task that is mainly performed to improve the efficiency of the machine learning process. Various tokenization methods have been proposed for application in the field of natural language processing, but studies have primarily focused on efficiently segmenting text. Few studies have been conducted on the Korean language to explore what tokenization methods are suitable for document classification task. In this paper, an exploratory study was performed to find the most suitable tokenization method to improve the performance of a representative spatio-temporal document classifier in Korean. For the experiment, a convolutional neural network model was used, and for the final performance comparison, tasks were selected for document classification where performance largely depends on the tokenization method. As a tokenization method for comparative experiments, commonly used Jamo, Character, and Word units were adopted. As a result of the experiment, it was confirmed that the tokenization of word units showed excellent performance in the case of representative spatio-temporal document classification task where the semantic embedding ability of the token itself is important.

• ETRI Journal
• /
• v.39 no.4
• /
• pp.455-466
• /
• 2017

This study focuses on a method for sequential data augmentation in order to alleviate data sparseness problems. Specifically, we present corpus expansion techniques for enhancing the coverage of a language model. Recent recurrent neural network studies show that a seq2seq model can be applied for addressing language generation issues; it has the ability to generate new sentences from given input sentences. We present a method of corpus expansion using a sentence-chain based seq2seq model. For training the seq2seq model, sentence chains are used as triples. The first two sentences in a triple are used for the encoder of the seq2seq model, while the last sentence becomes a target sequence for the decoder. Using only internal resources, evaluation results show an improvement of approximately 7.6% relative perplexity over a baseline language model of Korean text. Additionally, from a comparison with a previous study, the sentence chain approach reduces the size of the training data by 38.4% while generating 1.4-times the number of n-grams with superior performance for English text.

• MALSORI
• /
• no.56
• /
• pp.207-223
• /
• 2005

The goal of language model adaptation is to improve the background language model with a relatively small adaptation corpus. This study presents a language model adaptation technique where additional text data for the adaptation do not exist. We propose the information retrieval (IR) technique with N-gram language modeling to collect the adaptation corpus from baseline text data. We also propose to use a dynamic language model interpolation coefficient to combine the background language model and the adapted language model. The interpolation coefficient is estimated from the word hypotheses obtained by segmenting the input speech data reserved for held-out validation data. This allows the final adapted model to improve the performance of the background model consistently The proposed approach reduces the word error rate by $13.6\%$ relative to baseline 4-gram for two-hour broadcast news speech recognition.