• Journal of Intelligence and Information Systems
• /
• v.23 no.2
• /
• pp.71-88
• /
• 2017

Language models were originally developed for speech recognition and language processing. Using a set of example sentences, a language model predicts the next word or character based on sequential input data. N-gram models have been widely used but this model cannot model the correlation between the input units efficiently since it is a probabilistic model which are based on the frequency of each unit in the training set. Recently, as the deep learning algorithm has been developed, a recurrent neural network (RNN) model and a long short-term memory (LSTM) model have been widely used for the neural language model (Ahn, 2016; Kim et al., 2016; Lee et al., 2016). These models can reflect dependency between the objects that are entered sequentially into the model (Gers and Schmidhuber, 2001; Mikolov et al., 2010; Sundermeyer et al., 2012). In order to learning the neural language model, texts need to be decomposed into words or morphemes. Since, however, a training set of sentences includes a huge number of words or morphemes in general, the size of dictionary is very large and so it increases model complexity. In addition, word-level or morpheme-level models are able to generate vocabularies only which are contained in the training set. Furthermore, with highly morphological languages such as Turkish, Hungarian, Russian, Finnish or Korean, morpheme analyzers have more chance to cause errors in decomposition process (Lankinen et al., 2016). Therefore, this paper proposes a phoneme-level language model for Korean language based on LSTM models. A phoneme such as a vowel or a consonant is the smallest unit that comprises Korean texts. We construct the language model using three or four LSTM layers. Each model was trained using Stochastic Gradient Algorithm and more advanced optimization algorithms such as Adagrad, RMSprop, Adadelta, Adam, Adamax, and Nadam. Simulation study was done with Old Testament texts using a deep learning package Keras based the Theano. After pre-processing the texts, the dataset included 74 of unique characters including vowels, consonants, and punctuation marks. Then we constructed an input vector with 20 consecutive characters and an output with a following 21st character. Finally, total 1,023,411 sets of input-output vectors were included in the dataset and we divided them into training, validation, testsets with proportion 70:15:15. All the simulation were conducted on a system equipped with an Intel Xeon CPU (16 cores) and a NVIDIA GeForce GTX 1080 GPU. We compared the loss function evaluated for the validation set, the perplexity evaluated for the test set, and the time to be taken for training each model. As a result, all the optimization algorithms but the stochastic gradient algorithm showed similar validation loss and perplexity, which are clearly superior to those of the stochastic gradient algorithm. The stochastic gradient algorithm took the longest time to be trained for both 3- and 4-LSTM models. On average, the 4-LSTM layer model took 69% longer training time than the 3-LSTM layer model. However, the validation loss and perplexity were not improved significantly or became even worse for specific conditions. On the other hand, when comparing the automatically generated sentences, the 4-LSTM layer model tended to generate the sentences which are closer to the natural language than the 3-LSTM model. Although there were slight differences in the completeness of the generated sentences between the models, the sentence generation performance was quite satisfactory in any simulation conditions: they generated only legitimate Korean letters and the use of postposition and the conjugation of verbs were almost perfect in the sense of grammar. The results of this study are expected to be widely used for the processing of Korean language in the field of language processing and speech recognition, which are the basis of artificial intelligence systems.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.63-83
• /
• 2019

Investors prefer to look for trading points based on the graph shown in the chart rather than complex analysis, such as corporate intrinsic value analysis and technical auxiliary index analysis. However, the pattern analysis technique is difficult and computerized less than the needs of users. In recent years, there have been many cases of studying stock price patterns using various machine learning techniques including neural networks in the field of artificial intelligence(AI). In particular, the development of IT technology has made it easier to analyze a huge number of chart data to find patterns that can predict stock prices. Although short-term forecasting power of prices has increased in terms of performance so far, long-term forecasting power is limited and is used in short-term trading rather than long-term investment. Other studies have focused on mechanically and accurately identifying patterns that were not recognized by past technology, but it can be vulnerable in practical areas because it is a separate matter whether the patterns found are suitable for trading. When they find a meaningful pattern, they find a point that matches the pattern. They then measure their performance after n days, assuming that they have bought at that point in time. Since this approach is to calculate virtual revenues, there can be many disparities with reality. The existing research method tries to find a pattern with stock price prediction power, but this study proposes to define the patterns first and to trade when the pattern with high success probability appears. The M & W wave pattern published by Merrill(1980) is simple because we can distinguish it by five turning points. Despite the report that some patterns have price predictability, there were no performance reports used in the actual market. The simplicity of a pattern consisting of five turning points has the advantage of reducing the cost of increasing pattern recognition accuracy. In this study, 16 patterns of up conversion and 16 patterns of down conversion are reclassified into ten groups so that they can be easily implemented by the system. Only one pattern with high success rate per group is selected for trading. Patterns that had a high probability of success in the past are likely to succeed in the future. So we trade when such a pattern occurs. It is a real situation because it is measured assuming that both the buy and sell have been executed. We tested three ways to calculate the turning point. The first method, the minimum change rate zig-zag method, removes price movements below a certain percentage and calculates the vertex. In the second method, high-low line zig-zag, the high price that meets the n-day high price line is calculated at the peak price, and the low price that meets the n-day low price line is calculated at the valley price. In the third method, the swing wave method, the high price in the center higher than n high prices on the left and right is calculated as the peak price. If the central low price is lower than the n low price on the left and right, it is calculated as valley price. The swing wave method was superior to the other methods in the test results. It is interpreted that the transaction after checking the completion of the pattern is more effective than the transaction in the unfinished state of the pattern. Genetic algorithms(GA) were the most suitable solution, although it was virtually impossible to find patterns with high success rates because the number of cases was too large in this simulation. We also performed the simulation using the Walk-forward Analysis(WFA) method, which tests the test section and the application section separately. So we were able to respond appropriately to market changes. In this study, we optimize the stock portfolio because there is a risk of over-optimized if we implement the variable optimality for each individual stock. Therefore, we selected the number of constituent stocks as 20 to increase the effect of diversified investment while avoiding optimization. We tested the KOSPI market by dividing it into six categories. In the results, the portfolio of small cap stock was the most successful and the high vol stock portfolio was the second best. This shows that patterns need to have some price volatility in order for patterns to be shaped, but volatility is not the best.

• Journal of Intelligence and Information Systems
• /
• v.24 no.1
• /
• pp.205-225
• /
• 2018

Convolutional Neural Network (ConvNet) is one class of the powerful Deep Neural Network that can analyze and learn hierarchies of visual features. Originally, first neural network (Neocognitron) was introduced in the 80s. At that time, the neural network was not broadly used in both industry and academic field by cause of large-scale dataset shortage and low computational power. However, after a few decades later in 2012, Krizhevsky made a breakthrough on ILSVRC-12 visual recognition competition using Convolutional Neural Network. That breakthrough revived people interest in the neural network. The success of Convolutional Neural Network is achieved with two main factors. First of them is the emergence of advanced hardware (GPUs) for sufficient parallel computation. Second is the availability of large-scale datasets such as ImageNet (ILSVRC) dataset for training. Unfortunately, many new domains are bottlenecked by these factors. For most domains, it is difficult and requires lots of effort to gather large-scale dataset to train a ConvNet. Moreover, even if we have a large-scale dataset, training ConvNet from scratch is required expensive resource and time-consuming. These two obstacles can be solved by using transfer learning. Transfer learning is a method for transferring the knowledge from a source domain to new domain. There are two major Transfer learning cases. First one is ConvNet as fixed feature extractor, and the second one is Fine-tune the ConvNet on a new dataset. In the first case, using pre-trained ConvNet (such as on ImageNet) to compute feed-forward activations of the image into the ConvNet and extract activation features from specific layers. In the second case, replacing and retraining the ConvNet classifier on the new dataset, then fine-tune the weights of the pre-trained network with the backpropagation. In this paper, we focus on using multiple ConvNet layers as a fixed feature extractor only. However, applying features with high dimensional complexity that is directly extracted from multiple ConvNet layers is still a challenging problem. We observe that features extracted from multiple ConvNet layers address the different characteristics of the image which means better representation could be obtained by finding the optimal combination of multiple ConvNet layers. Based on that observation, we propose to employ multiple ConvNet layer representations for transfer learning instead of a single ConvNet layer representation. Overall, our primary pipeline has three steps. Firstly, images from target task are given as input to ConvNet, then that image will be feed-forwarded into pre-trained AlexNet, and the activation features from three fully connected convolutional layers are extracted. Secondly, activation features of three ConvNet layers are concatenated to obtain multiple ConvNet layers representation because it will gain more information about an image. When three fully connected layer features concatenated, the occurring image representation would have 9192 (4096+4096+1000) dimension features. However, features extracted from multiple ConvNet layers are redundant and noisy since they are extracted from the same ConvNet. Thus, a third step, we will use Principal Component Analysis (PCA) to select salient features before the training phase. When salient features are obtained, the classifier can classify image more accurately, and the performance of transfer learning can be improved. To evaluate proposed method, experiments are conducted in three standard datasets (Caltech-256, VOC07, and SUN397) to compare multiple ConvNet layer representations against single ConvNet layer representation by using PCA for feature selection and dimension reduction. Our experiments demonstrated the importance of feature selection for multiple ConvNet layer representation. Moreover, our proposed approach achieved 75.6% accuracy compared to 73.9% accuracy achieved by FC7 layer on the Caltech-256 dataset, 73.1% accuracy compared to 69.2% accuracy achieved by FC8 layer on the VOC07 dataset, 52.2% accuracy compared to 48.7% accuracy achieved by FC7 layer on the SUN397 dataset. We also showed that our proposed approach achieved superior performance, 2.8%, 2.1% and 3.1% accuracy improvement on Caltech-256, VOC07, and SUN397 dataset respectively compare to existing work.

• Journal of Internet Computing and Services
• /
• v.14 no.5
• /
• pp.1-10
• /
• 2013

RFID(Radio Frequency Identification) system is non-contact identification technology. A basic RFID system consists of a reader, and a set of tags. RFID tags can be divided into active and passive tags. Active tags with power source allows their own operation execution and passive tags are small and low-cost. So passive tags are more suitable for distribution industry than active tags. A reader processes the information receiving from tags. RFID system achieves a fast identification of multiple tags using radio frequency. RFID systems has been applied into a variety of fields such as distribution, logistics, transportation, inventory management, access control, finance and etc. To encourage the introduction of RFID systems, several problems (price, size, power consumption, security) should be resolved. In this paper, we proposed an algorithm to significantly alleviate the collision problem caused by simultaneous responses of multiple tags. In the RFID systems, in anti-collision schemes, there are three methods: probabilistic, deterministic, and hybrid. In this paper, we introduce ALOHA-based protocol as a probabilistic method, and Tree-based protocol as a deterministic one. In Aloha-based protocols, time is divided into multiple slots. Tags randomly select their own IDs and transmit it. But Aloha-based protocol cannot guarantee that all tags are identified because they are probabilistic methods. In contrast, Tree-based protocols guarantee that a reader identifies all tags within the transmission range of the reader. In Tree-based protocols, a reader sends a query, and tags respond it with their own IDs. When a reader sends a query and two or more tags respond, a collision occurs. Then the reader makes and sends a new query. Frequent collisions make the identification performance degrade. Therefore, to identify tags quickly, it is necessary to reduce collisions efficiently. Each RFID tag has an ID of 96bit EPC(Electronic Product Code). The tags in a company or manufacturer have similar tag IDs with the same prefix. Unnecessary collisions occur while identifying multiple tags using Query Tree protocol. It results in growth of query-responses and idle time, which the identification time significantly increases. To solve this problem, Collision Tree protocol and M-ary Query Tree protocol have been proposed. However, in Collision Tree protocol and Query Tree protocol, only one bit is identified during one query-response. And, when similar tag IDs exist, M-ary Query Tree Protocol generates unnecessary query-responses. In this paper, we propose Adaptive M-ary Query Tree protocol that improves the identification performance using m-bit recognition, collision information of tag IDs, and prediction technique. We compare our proposed scheme with other Tree-based protocols under the same conditions. We show that our proposed scheme outperforms others in terms of identification time and identification efficiency.

• Journal of the Korean Society for Marine Environment & Energy
• /
• v.15 no.4
• /
• pp.337-348
• /
• 2012

The Yellow Sea Ecoregion Conservation Project is a joint international project which is carried out under the purposes of conserving the habitats in the Yellow Sea Ecoregion and biodiversity from various threats that damage ecosystems, informing the importance and value of the Yellow Sea Ecoregion, and promoting the understanding and interests of Korea, China and Japan. Subsequent to the Yellow Sea Ecoregion Planning Programme which had been performed during the period from 2002 to 2006, the Yellow Sea Ecoregion Support Project has been performed over 7 years in total from 2007 to 2014. Panasonic is sponsoring the financing of the project, and the organizations in charge of the project by country are Korea Institute of Ocean Science & Technology for Korea and World Wide Fund for Nature branches for China and Japan. While the Yellow Sea Ecoregion Planning Programme was focused on the biological assessment and the selection of potential priority area by scientific review, the Yellow Sea Ecoregion Support Project is oriented toward practical activities targeting more diversified stakeholder. Especially, this project plans to support direct conservation activities in the region and participation and cooperation from local residents. The project plan is comprised of 3 phases. During the first period from 2008 to 2009, small grant projects were operated targeting 16 institutions of Korea and China, and for the second period from 2010 to 2012, one place each was selected as demonstration site for habitat conservation in Korea and China respectively and supported for three years to introduce the conservation method based on international standards such as the management of marine protected areas, ecosystem-based management and community-based management and simultaneously to develop habitat conservation activities suitable for national and regional characteristics. During the period from 2013 to 2014 which is the last phase, the project plans to distribute the performance of small grant projects and demonstration site activities through a series of forums among stakeholder. Through the activities described above, the recognition of general public on the conservation of the Yellow Sea Ecoregion was changed positively, and community-based management began to be reflected in the policies for the marine protected areas of central and local government.

• Journal of Internet Computing and Services
• /
• v.21 no.4
• /
• pp.17-23
• /
• 2020

Biometric information indicating measurement items related to human characteristics has attracted great attention as security technology with high reliability since there is no fear of theft or loss. Among these biometric information, fingerprints are mainly used in fields such as identity verification and identification. If there is a problem such as a wound, wrinkle, or moisture that is difficult to authenticate to the fingerprint image when identifying the identity, the fingerprint expert can identify the problem with the fingerprint directly through the preprocessing step, and apply the image processing algorithm appropriate to the problem. Solve the problem. In this case, by implementing artificial intelligence software that distinguishes fingerprint images with cuts and wrinkles on the fingerprint, it is easy to check whether there are cuts or wrinkles, and by selecting an appropriate algorithm, the fingerprint image can be easily improved. In this study, we developed a total of 17,080 fingerprint databases by acquiring all finger prints of 1,010 students from the Royal University of Cambodia, 600 Sokoto open data sets, and 98 Korean students. In order to determine if there are any injuries or wrinkles in the built database, criteria were established, and the data were validated by experts. The training and test datasets consisted of Cambodian data and Sokoto data, and the ratio was set to 8: 2. The data of 98 Korean students were set up as a validation data set. Using the constructed data set, five CNN-based architectures such as Classic CNN, AlexNet, VGG-16, Resnet50, and Yolo v3 were implemented. A study was conducted to find the model that performed best on the readings. Among the five architectures, ResNet50 showed the best performance with 81.51%.

• Journal of KIISE:Software and Applications
• /
• v.30 no.9
• /
• pp.829-842
• /
• 2003

In this Paper, we propose the new face detection and tracking method based on template matching for real-time applications such as, teleconference, telecommunication, front stage of surveillance system using face recognition, and video-phone applications. Since the main purpose of paper is to track a face regardless of various environments, we use template-based face tracking method. To generate robust face templates, we apply wavelet transform to the average face image and extract three types of wavelet template from transformed low-resolution average face. However template matching is generally sensitive to the change of illumination conditions, we apply Min-max normalization with histogram equalization according to the variation of intensity. Tracking method is also applied to reduce the computation time and predict precise face candidate region. Finally, facial components are also detected and from the relative distance of two eyes, we estimate the size of facial ellipse.

• Journal of Intelligence and Information Systems
• /
• v.21 no.2
• /
• pp.113-129
• /
• 2015

Location branding is a very important income making activity, by giving special meanings to a specific location while producing identity and communal value which are based around the understanding of a place's location branding concept methodology. Many other areas, such as marketing, architecture, and city construction, exert an influence creating an impressive brand image. A place brand which shows great recognition to both native people of S. Korea and foreigners creates significant economic effects. There has been research on creating a strategically and detailed place brand image, and the representative research has been carried out by Anholt who surveyed two million people from 50 different countries. However, the investigation, including survey research, required a great deal of effort from the workforce and required significant expense. As a result, there is a need to make more affordable, objective and effective research methods. The purpose of this paper is to find a way to measure the intensity of the image of the brand objective and at a low cost through text mining purposes. The proposed method extracts the keyword and the factors constructing the location brand image from the related web documents. In this way, we can measure the brand image intensity of the specific location. The performance of the proposed methodology was verified through comparison with Anholt's 50 city image consistency index ranking around the world. Four methods are applied to the test. First, RNADOM method artificially ranks the cities included in the experiment. HUMAN method firstly makes a questionnaire and selects 9 volunteers who are well acquainted with brand management and at the same time cities to evaluate. Then they are requested to rank the cities and compared with the Anholt's evaluation results. TM method applies the proposed method to evaluate the cities with all evaluation criteria. TM-LEARN, which is the extended method of TM, selects significant evaluation items from the items in every criterion. Then the method evaluates the cities with all selected evaluation criteria. RMSE is used to as a metric to compare the evaluation results. Experimental results suggested by this paper's methodology are as follows: Firstly, compared to the evaluation method that targets ordinary people, this method appeared to be more accurate. Secondly, compared to the traditional survey method, the time and the cost are much less because in this research we used automated means. Thirdly, this proposed methodology is very timely because it can be evaluated from time to time. Fourthly, compared to Anholt's method which evaluated only for an already specified city, this proposed methodology is applicable to any location. Finally, this proposed methodology has a relatively high objectivity because our research was conducted based on open source data. As a result, our city image evaluation text mining approach has found validity in terms of accuracy, cost-effectiveness, timeliness, scalability, and reliability. The proposed method provides managers with clear guidelines regarding brand management in public and private sectors. As public sectors such as local officers, the proposed method could be used to formulate strategies and enhance the image of their places in an efficient manner. Rather than conducting heavy questionnaires, the local officers could monitor the current place image very shortly a priori, than may make decisions to go over the formal place image test only if the evaluation results from the proposed method are not ordinary no matter what the results indicate opportunity or threat to the place. Moreover, with co-using the morphological analysis, extracting meaningful facets of place brand from text, sentiment analysis and more with the proposed method, marketing strategy planners or civil engineering professionals may obtain deeper and more abundant insights for better place rand images. In the future, a prototype system will be implemented to show the feasibility of the idea proposed in this paper.

• Asia-Pacific Journal of Business Venturing and Entrepreneurship
• /
• v.9 no.2
• /
• pp.1-13
• /
• 2014

This study is preliminary founding start-ups of restaurant entrepreneurs to study the characteristics of management in order to provide useful information was studied. Survey period is from 2013 March 15 to August 31. About the nature of the preparation and establishment founder findings restaurant entrepreneurs of Western Gyeongnam area were most often in the form of an independent establishment has been compiled, work experience and other industries operating experience was more than restaurant entrepreneurs of operating experience. Hypothesis Testing in accordance with results of this study are as follows. First, the 50 founders than 30 founders quickly when the customer complaint or that service, order food and non-food note is issued after obtaining the order or that provide food, no customer is often inconvenient check the sharing services, such as personnel activities were devoting a lot of effort. Second, the re-startups restaurant entrepreneurs than new startups restaurant entrepreneurs was founded after the founder of career-related customer complaints about food more active coping was, and re-startups restaurant entrepreneurs by the founder other than business founded by the founder of the food-related customer complaints more aggressively for coping, respectively. Third, restaurant entrepreneurs of the store operations management has integrity, words and actions match, such as the degree of belief in the promise of reliability and the possibility of failure, which means the degree of recognition and response efforts are having an impact deal. Fourth, restaurant of food service management services and after-sales service has impact on the founder of the self-efficacy and self-efficacy of pre-service features and reliability founder affecting. Fifth, the revenue of the restaurant for dealing with customer complaints management includes efforts are having an impact. Sixth, restaurant founder of reliability and customer care has influenced the self-efficacy. Seventh, management of operational management activities have a positive impact on business performance are.

• Asia pacific journal of information systems
• /
• v.19 no.2
• /
• pp.179-193
• /
• 2009

As the common saying attests, a publish-or-perish world, publishing is absolutely critical for academic researchers' successful careers. It is the most objectively-accepted academic performance criteria and the most viable way to attain public and academic recognition. Asia Pacific Journal of Information Systems(APJIS) has been recognized as the most influential domestic journal in Korean MIS field since July, 1991. Therefore, publishing in APJIS means your research is original, valid, and contributive. While most researchers learn how to publish an article in APJIS through a repetitive review process, thereby improving their chance of the' accepted' through their personal trial and error experiences, such valuable lessons and know-how tend to be kept personally and rarely shared. However, useful insights into research and publication skills could be also gained from sharing others' errors, neglect, and misjudgments which are equally critical in improving researchers' knowledge in the field (Murthy and Wiggins, 2002). For this reason, other academic disciplines make systematic efforts to examine the paper review process of major journals and share the findings from these studies with the rest of the research community members (Beyer et al., 1995; Cummings et al, 1985; Daft, 1995; Jauch and Wall, 1989; Murthy and Wiggins, 2002). Recognizing the urgent need to provide such type of information to MIS research community in Korea, we have chosen the most influential academic journal, APJIS with an intention to share the answer to the following research question: "What are the common problems found in the manuscripts either 'rejected' or 'required a major revision' by APJIS reviewers?" This study analyzes the review results of manuscripts submitted to APJIS (from January, 2006 to October, 2008), particularly those that were 'rejected' or required a 'major revision' at the first round. Based on Daft's(1995) study, twelve most-likelihood problems were defined and used to analyze the reviews. The twelve criteria for classification, or "twelve problems", are as follows: No theory, Concepts and operationalization not in alignment, Insufficient definition--theory, Insufficient rationale--design, Macrostructure--organization and flow, Amateur style and tone, Inadequate research design, Not relevant to the field, Overengineering, Conclusions not in alignment, Cutting up the data, and Poor editorial practice. Upon the approval of the editorial board of APJIS, the total 252 reviews, including 11 cases of 2005 and 241 cases from July, 2006 to October, 2008, were received without any information about manuscripts, authors, or reviewers. Eleven cases of 2005 were used in the pilot test because the data of 2005 were not in complete enumeration, and the 241 reviews (113 cases of 'rejection' and 128 ones of 'major revision') of 2006, 2007, and 2008 were examined in this study. Our findings show that insufficient rationale-design(20.25%), no theory(18.45%), and insufficient definition--theory(15.69%) were the three leading reasons of 'rejection' and 'major revision.' Between these two results, the former followed the same order of three major reasons as an overall analysis (insufficient rationale-design, no theory, and insufficient definition-theory), but the latter followed the order of insufficient rationale--design, insufficient definition--theory, and no theory. Using Daft's three major skills-- 'theory skills', 'design skills', and 'communication skills'-- twelve criteria were reclassified into 'theory problems', 'design problems', and 'communication problems' to derive more practical implications of our findings. Our findings show that 'theory problems' occupied 43.48%, 'design problems' were 30.86%, and 'communication problems' were 25.86%. In general, the APJIS reviewers weigh each of these three problem areas almost equally. Comparing to other disciplines like management field shown in Daft's study, the portion of 'design problems' and 'communication problems' are much higher in manuscripts submitted to the APJIS than in those of Administrative Science Quarterly and Academy of Management Journal even though 'theory problems' are the most predominant in both disciplines.