• Smart Media Journal
• /
• v.10 no.1
• /
• pp.39-46
• /
• 2021

This study is about a method of extracting a summary from a news article in consideration of the importance of each sentence constituting the article. We propose a method of calculating sentence importance by extracting the probabilities of topic sentence, similarity with article title and other sentences, and sentence position as characteristics that affect sentence importance. At this time, a hypothesis is established that the Topic Sentence will have a characteristic distinct from the general sentence, and a deep learning-based classification model is trained to obtain a topic sentence probability value for the input sentence. Also, using the pre-learned ELMo language model, the similarity between sentences is calculated based on the sentence vector value reflecting the context information and extracted as sentence characteristics. The topic sentence classification performance of the LSTM and BERT models was 93% accurate, 96.22% recall, and 89.5% precision, resulting in high analysis results. As a result of calculating the importance of each sentence by combining the extracted sentence characteristics, it was confirmed that the performance of extracting the topic sentence was improved by about 10% compared to the existing TextRank algorithm.

• Journal of Intelligence and Information Systems
• /
• v.14 no.2
• /
• pp.121-136
• /
• 2008

Although domestic research funding organizations require patented technologies as an outcome of financial aids, they have much difficulty in evaluating qualitative value of the patented technology due to lack of systematic methods. Especially, because citation data is not essential to patent application in Korea, it is very difficult to evaluate a patent using the incomplete citation data. This study proposes a method for evaluating technology importance of a patent when there is no or insufficient citation data in patents. The technology importance of a patent can be evaluated objectively and quantitatively by the proposed method which consists of 5 steps such as selection of a target patent, collection of related patents, preparation of key word vector, clustering patents, and technological importance assessment. The method was applied to a patent on 'user identification method for payment using mobile terminal' in order to evaluate technology importance and demonstrate how the method works.

• KIPS Transactions on Software and Data Engineering
• /
• v.4 no.6
• /
• pp.261-268
• /
• 2015

In spoken chatting systems, users'spoken queries are converted to text queries using automatic speech recognition (ASR) engines. If the top-1 results of the ASR engines are incorrect, these errors are propagated to the spoken chatting systems. To improve the top-1 accuracies of ASR engines, we propose a post-processing model to rearrange the top-n outputs of ASR engines using a ranking support vector machine (RankSVM). On the other hand, a number of chatting sentences are needed to train chatting systems. If new chatting sentences are not frequently added to training data, responses of the chatting systems will be old-fashioned soon. To resolve this problem, we propose a data collection model to automatically select chatting sentences from TV and movie scenarios using a support vector machine (SVM). In the experiments, the post-processing model showed a higher precision of 4.4% and a higher recall rate of 6.4% compared to the baseline model (without post-processing). Then, the data collection model showed the high precision of 98.95% and the recall rate of 57.14%.

• Journal of Korean Society for Quality Management
• /
• v.45 no.1
• /
• pp.117-128
• /
• 2017

Purpose: This study analyzes automobile quality review data to develop alternative analytical method of informal data. Existing methods to analyze informal data are based mainly on the frequency of informal data, however, this research tries to use correlation information of each informal data. Method: After sentimental analysis to acquire the user information for automobile products, three classification methods, that is, $na{\ddot{i}}ve$ Bayes, random forest, and support vector machine, were employed to accurately classify the informal user opinions with respect to automobile qualities. Additionally, Word2vec was applied to discover correlated information about informal data. Result: As applicative results of three classification methods, random forest method shows most effective results compared to the other classification methods. Word2vec method manages to discover closest relevant data with automobile components. Conclusion: The proposed method shows its effectiveness in terms of accuracy and sensitivity on the analysis of informal quality data, however, only two sentiments (positive or negative) can be categorized due to human errors. Further studies are required to derive more sentiments to accurately classify informal quality data. Word2vec method also shows comparative results to discover the relevance of components precisely.

• Journal of KIISE:Software and Applications
• /
• v.29 no.10
• /
• pp.684-693
• /
• 2002

Electronic Ink is a stored data in the form of the handwritten text or the script without converting it into ASCII by handwritten recognition on the pen-based computers and Personal Digital Assistants(PDAs) for supporting natural and convenient data input. One of the most Important issue is to search the electronic ink in order to use it. We proposed and implemented a script matching algorithm for the electronic ink. Proposed matching algorithm separated the input stroke into a set of primitive stroke using the curvature of the stroke curve. After determining the type of separated strokes, it produced a stroke feature vector. And then it calculated the distance between the stroke feature vector of input strokes and one of strokes in the database using the dynamic programming technique. We did various experiments and our algorithm showed high matching rate over 97.7% for only the Korean script and 94% for the data mixed Korean with the Chinese character.

• Journal of the Korean Data and Information Science Society
• /
• v.21 no.4
• /
• pp.689-697
• /
• 2010

In this study, we propose a new procedure for improving the character recognition of text area extracted from video images. The recognition of strings extracted from video, which are mixed with Hangul, English, numbers and special characters, etc., is more difficult than general character recognition because of various fonts and size, graphic forms of letters tilted image, disconnection, miscellaneous videos, tangency, characters of low definition, etc. We improved the recognition rate by taking commonly used letters and leaving out the barely used ones instead of recognizing all of the letters, and then using SVM and RDA character recognition methods. Our numerical results indicate that combining SVM and RDA performs better than other methods.

• International journal of advanced smart convergence
• /
• v.5 no.2
• /
• pp.53-58
• /
• 2016

With the growing use of big data and data mining, it serves to understand how such techniques can be used to understand various relationships in the healthcare field. This study uses hierarchical methods of data analysis to explore similarities in hospitalization across several New York state counties. The study utilized methods of measuring crowding distance of data for age-specific hospitalization period. Crowding distance is defined as the longest distance, or least similarity, between urban cities. It is expected that the city of Clinton have the greatest distance, while Albany the other cities are closer because they are connected by the shortest distance to each step. Similarities were stronger across hospital stays categorized by age. Hierarchical clustering can be applied to predict the similarity of data across the 10 cities of hospitalization with the measurement of crowding distance. In order to enhance the performance of hierarchical clustering, comparison can be made across congestion distance when crowding distance is applied first through the application of converting text to an attribute vector. Measurements of similarity between two objects are dependent on the measurement method used in clustering but is distinguished from the similarity of the distance; where the smaller the distance value the more similar two things are to one other. By applying this specific technique, it is found that the distance between crowding is reduced consistently in relationship to similarity between the data increases to enhance the performance of the experiments through the application of special techniques. Furthermore, through the similarity by city hospitalization period, when the construction of hospital wards in cities, by referring to results of experiments, or predict possible will land to the extent of the size of the hospital facilities hospital stay is expected to be useful in efficiently managing the patient in a similar area.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.5
• /
• pp.974-983
• /
• 2017

In this paper, we have discussed the search function that can be embedded and used on Android-based applications. We used BM25 to suppress insignificant and too frequent words such as postpositions, Pivoted Length Normalization technique used to resolve the search priority problem related to each item's length, and Rocchio's method to pull items inferred to be related to the query closer to the query vector on Vector Space Model to support implicit feedback function. The index operation is divided into two methods; simple index to support offline operation and complex index for online operation. The implementation uses query inference function to guess user's future input by collating given present input with indexed data and with it the function is able to handle and correct user's error. Thus the implementation could be easily adopted into smartphone applications to improve their search functions.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.30-39
• /
• 2007

The existing content-based spam mail filtering algorithms have difficulties in filtering spam mails when e-mails contain images but little text. In this thesis we propose an efficient spam mail classification algorithm that utilizes the link structure of e-mails. We compute the number of hyperlinks in an e-mail and the in-link frequencies of the web pages hyperlinked in the e-mail. Using these two features we classify spam mails and legitimate mails based on the decision tree trained for spam mail classification. We also suggest a hybrid system combining three different algorithms by majority voting: the link structure analysis algorithm, a modified link structure analysis algorithm, in which only the host part of the hyperlinked pages of an e-mail is used for link structure analysis, and the content-based method using SVM (support vector machines). The experimental results show that the link structure analysis algorithm slightly outperforms the existing content-based method with the accuracy of 94.8%. Moreover, the hybrid system achieves the accuracy of 97.0%, which is a significant performance improvement over the existing method.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.12 no.2
• /
• pp.281-286
• /
• 2017

SNS is a platformwhere users build a social relationship and share opinions and informations. To do these, it supports a text, an image and a video data. As it is possible to exploit the location data of a smart device, SNS tries to use the location data. However,since SNS does not support the coordinate data, it provides the restricted function sharing the image map instead of the vector map. In this paper, we propose a chatting systemsharing pathsto support the coordinate data on a classical SNS. On the proposed system, it is possible for usersjoined in a roomto watch a vector map of same area and exchange texts. If a user builds a path on the map, the system propagates the coordinate data of the generated path and the other users joined in the room watch the path immediately. The implemented chatting systemhasthe benefit to share the information related a map between users using coordinate data.