• Journal of Intelligence and Information Systems
• /
• v.26 no.1
• /
• pp.1-21
• /
• 2020

With the rapid development of artificial intelligence technology, various techniques have been developed to extract meaningful information from unstructured text data which constitutes a large portion of big data. Over the past decades, text mining technologies have been utilized in various industries for practical applications. In the field of business intelligence, it has been employed to discover new market and/or technology opportunities and support rational decision making of business participants. The market information such as market size, market growth rate, and market share is essential for setting companies' business strategies. There has been a continuous demand in various fields for specific product level-market information. However, the information has been generally provided at industry level or broad categories based on classification standards, making it difficult to obtain specific and proper information. In this regard, we propose a new methodology that can estimate the market sizes of product groups at more detailed levels than that of previously offered. We applied Word2Vec algorithm, a neural network based semantic word embedding model, to enable automatic market size estimation from individual companies' product information in a bottom-up manner. The overall process is as follows: First, the data related to product information is collected, refined, and restructured into suitable form for applying Word2Vec model. Next, the preprocessed data is embedded into vector space by Word2Vec and then the product groups are derived by extracting similar products names based on cosine similarity calculation. Finally, the sales data on the extracted products is summated to estimate the market size of the product groups. As an experimental data, text data of product names from Statistics Korea's microdata (345,103 cases) were mapped in multidimensional vector space by Word2Vec training. We performed parameters optimization for training and then applied vector dimension of 300 and window size of 15 as optimized parameters for further experiments. We employed index words of Korean Standard Industry Classification (KSIC) as a product name dataset to more efficiently cluster product groups. The product names which are similar to KSIC indexes were extracted based on cosine similarity. The market size of extracted products as one product category was calculated from individual companies' sales data. The market sizes of 11,654 specific product lines were automatically estimated by the proposed model. For the performance verification, the results were compared with actual market size of some items. The Pearson's correlation coefficient was 0.513. Our approach has several advantages differing from the previous studies. First, text mining and machine learning techniques were applied for the first time on market size estimation, overcoming the limitations of traditional sampling based- or multiple assumption required-methods. In addition, the level of market category can be easily and efficiently adjusted according to the purpose of information use by changing cosine similarity threshold. Furthermore, it has a high potential of practical applications since it can resolve unmet needs for detailed market size information in public and private sectors. Specifically, it can be utilized in technology evaluation and technology commercialization support program conducted by governmental institutions, as well as business strategies consulting and market analysis report publishing by private firms. The limitation of our study is that the presented model needs to be improved in terms of accuracy and reliability. The semantic-based word embedding module can be advanced by giving a proper order in the preprocessed dataset or by combining another algorithm such as Jaccard similarity with Word2Vec. Also, the methods of product group clustering can be changed to other types of unsupervised machine learning algorithm. Our group is currently working on subsequent studies and we expect that it can further improve the performance of the conceptually proposed basic model in this study.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.32 no.4D
• /
• pp.345-356
• /
• 2012

Road pavement requires repetitive maintenance works to maintain satisfactory service level to the public. However, the repetitive maintenance works upon deteriorated pavement structure make negative effects to deterioration speed. It often leads to inefficient use of limited budget. For that reason, the pavements require reconstruction work to recover their original performance. Recently, construction demands in the Korean national highway have already been reached to maximum level, and the aged pavements start to demand much more reconstruction works. However, in the real world, road agencies have often been confused when they determine maintenance design for such aged road sections due to budget constraint. It is because there is no reliable long-term maintenance strategy that supports their decision making. To support their decision making, this paper aimed to suggest the best maintenance strategy considering changing process of pavement performance by repetitive maintenance works. As an analysis method, probability distribution and hazard function to estimate the life expectancy were adopted, and then the results were used for long-term life cycle cost analysis with deterministic or Monte-Carlo method under various scenarios. As an empirical study, the Korean national highway data that has long-maintenance history data since 1986 has been applied. Last, this paper considered quality assurance of maintenance work to improve maintenance quality. These could be important information as a part of long-term maintenance strategy of pavement.

• Journal of Bio-Environment Control
• /
• v.21 no.3
• /
• pp.286-293
• /
• 2012

This study was initiated to find out the suitable perennial ground covers naturally grown in thirteen organic orchards in Chonnam Province as a organic nutrient source for maintaining annual fruit tree growth. The ground covers were observed in April, June, and August in the orchards. Agropyron tsukusinense and Panicum virgatum observed in April and June, respectively, produced the highest dry weight, which increased amounts of N, $P_2O_5$, and $K_2O$, mineralizing from the residue in the ground covers. The occurrence of perennial ground covers in August decreased compared to April and June. Amount of residue in mowed Agropyron tsukusinense and Panicum virgatum satisfied nutrient demand (N; 20 kg/10a, $P_2O_5$; 11 kg/10a, and $K_2O$; 19 kg/10a) to achieve the annual growth of twenty-year old fruit tree.

• Environmental and Resource Economics Review
• /
• v.22 no.4
• /
• pp.613-641
• /
• 2013

This paper uses the error correction model to analyse dynamic gasoline price adjustments of the four refiners. Unlike the existing studies, this model allows a refiner's asymmetric adjustment to changes in the other refiners' prices as well as in its own price and costs. With the estimation results, we can obtain the following findings. First, there are the asymmetric price adjustments to changes in exchange rate and international gasoline price, but showing opposing directions. Second, for most of the refiners, the prices respond immediately to the lagged deviation from the long run equilibrium price, but asymmetrically respond for a few refiners. Third, there are some refiners that adjust their price to the other refiners' price deviation from the long run equilibrium. For some refiners, there are competitive price adjustments to the others' price deviations. These findings imply that a refiner faces inelastic demand, intends to maintain implicitly a relative level of its own price to others, and tends to respond competitively to the others' price deviation from the equilibrium.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.32 no.1A
• /
• pp.1-10
• /
• 2012

Along with the development of coasts, islands and mountains, the demand of long-span bridges increases which, in turn, brings forth the construction of cable-supported bridges like suspension and cable-stayed bridges. There are various types of statically indeterminate structures widely applied that supported the main girder with stay cables, main cables, hanger cables with aesthetic structural appearance. As to the cable-supported bridges, the health monitoring of a bridge can be identified by measuring tension force on cable repeatedly. The tension force on cable is measured either by direct measurement of stress of cable using load cell or hydraulic jack, or by vibration method estimating tension force using cable shape and measured dynamic characteristics. In this study, a method to estimate dynamic characteristics of hanger cables by using a digital image processing is suggested. Digital images are acquired by a portable digital camcorder, which is the sensor to remotely measure dynamic responses considering convenient and economical aspects for use. A digital image correlation(DIC) technique is applied for digital image processing, and an image transform function(ITF) to correct the geometric distortion induced from the deformed images is used to estimate subpixel. And, the correction of motion of vision-based measurement system using a fixed object in an image without installing additional sensor can be enhanced the resolution of dynamic responses and modal frequencies of hanger cables.

• Journal of Korean Society of Environmental Engineers
• /
• v.35 no.7
• /
• pp.464-471
• /
• 2013

In this study, the resource efficiency and future metal resource requirement in photovoltaic (PV) production system were evaluated by using material balance data and life cycle assesment (LCA) method. As a result, in the resource efficiency of ferrous and non-ferrous metal, lead and tin had higher resource efficiency than other materials in all PV systems (SC-Si, MC-Si, CI(G)S, CdTe). In the resource efficiency of rare metals, gallium and rhenium in silicon system and rhenium and rhodium in thin-film system ranked as the first and second high resource efficiency. In case of rare earth metal, gadolinium and samarium took higher resource efficiency. The results of the future metal resource requirement in PV systems showed that 2,545,670 ton of aluminium, 92,069 ton of zinc, 22,044 ton of copper, 1,695 ton of tin and 31 ton of nickel will be needed by 2030 in South Korea, except resource recycling supplement.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.6
• /
• pp.212-217
• /
• 2020

The purpose of this study was to analyze the volatility and properties of a time series for tangerine prices in Jeju using the GARCH model of Bollerslev(1986). First, it was found that the time series for the rate of change in tangerine prices had a thicker tail rather than a normal distribution. At a significance level of 1%, the Jarque-Bera statistic led to a rejection of the null hypothesis that the distribution of the time series for the rate of change in tangerine prices is normally distributed. Second, the correlation between the time series was high based on the Ljung-Box Q statistic, which was statistically verified through the ARCH-LM test. Third, the results of the GARCH(1,1) model estimation showed statistically significant results at a significance level of 1%, except for the constant of the mean equation. The persistence parameter value of the variance equation was estimated to be close to 1, which means that there is a high possibility that a similar level of volatility will be present in the future. Finally, it is expected that the results of this study can be used as basic data to optimize the government's tangerine supply and demand control policy.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.6
• /
• pp.656-662
• /
• 2016

We have provided a variety of domestic construction technology related full-text services through the Construction Technology Digital Library system since 2001. CODIL is a system that services the database related to construction technology data. On the other hand, there is growing demand for DB every year, but the required budget is shrinking. Therefore, this study investigated the satisfaction to effectively service the construction technique-related full-text with a limited budget. The monetary value of full-text to express satisfaction with the quantified value was estimated using the Tobit model. The Tobit model is used as a contingent valuation method to estimate the value of non-market goods. This model is the limited dependent variable regression model to observations by censoring the limit of the left side or right side so that a biased outlier is not reflected in the willingness to pay. A survey was conducted by sampling 312 respondents. The mean, median, truncating the willingness of payment were calculated for the six types of the full-text services using the Tobit model. The statistically significant variables affecting the willingness to pay for the full-text services were identified. The mean value of per the full-text service was estimated to be 46,530 won. The significance of this study was to use the Tobit model to estimate the value of the construction technology-related full-text services for the first time in Korea.

• Journal of Korea Spatial Information System Society
• /
• v.10 no.4
• /
• pp.67-75
• /
• 2008

This study is for the effective training strategy of human resources for the u-City construction to support the u-City human resource development plan of the Ministry of Land, Transport and Maritime Affairs (MLTM). One of the biggest problems concerning u-City is the shortage of advanced human resources for the u-City constructions. The characteristic of u-City makes u-City related human resources had knowledge of various fields including IT, GIS, construction engineering, urban planning and so on. But, there are only a few programs to train u-City related human resources. Therefore, this research established the objective of the training strategy for u-City human resource development as "the training strategy of human resources for the successful u-City constructions". To achieve this objective, four different core strategies are established like followings: (1) demander-oriented education, (2) regional balanced education, (3) integration education of u-City related subjects, (4) u-City related education infrastructure development. These 4 different core strategies can be achieved from 5 sub projects like followings: (1) demand estimation of u-City human resources, (2) u-City education from selected regional core universities, (3) u-City education from u-City human resource education centers, (4) online education and (5) construction of education infrastructures. These 5 interrelated sub projects can be preconditions of the successful human resource strategy development.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.38 no.2
• /
• pp.138-148
• /
• 2001

Digital Video (DV) coding standards for digital video cassette recorder are based mainly on DCT and variable length coding. DV has low hardware complexity but high compressed bit rate of about 26 Mb/s. Thus, it is necessary to encode video with low complex video coding at the studios and then transcode compressed video into MPEG-2 for video-on-demand system. Because these coding methods exploit DCT, transcoding in the DCT domain can reduce computational complexity by excluding duplicated procedures. In transcoding DV into MPEC-2 intra coding, multiplying matrix by transformed data is used for 4:1:1-to-4:2:2 chroma format conversion and the conversion from 2-4-8 to 8-8 DCT mode, and therefore enables parallel processing. Variance of sub block for MPEG-2 rate control is computed completely in the DCT domain. These are verified through experiments. We estimate motion hierarchically using DCT coefficients for transcoding into MPEG-2 inter coding. First, we estimate motion of a macro block (MB) only with 4 DC values of 4 sub blocks and then estimate motion with 16-point MB using IDCT of 2$\times$2 low frequencies in each sub block, and finish estimation at a sub pixel as the fifth step. ME with overlapped search range shows better PSNR performance than ME without overlapping.