• Journal of Intelligence and Information Systems
• /
• v.24 no.2
• /
• pp.125-148
• /
• 2018

Recently, as the demand for big data analysis increases, cases of analyzing unstructured data and using the results are also increasing. Among the various types of unstructured data, text is used as a means of communicating information in almost all fields. In addition, many analysts are interested in the amount of data is very large and relatively easy to collect compared to other unstructured and structured data. Among the various text analysis applications, document classification which classifies documents into predetermined categories, topic modeling which extracts major topics from a large number of documents, sentimental analysis or opinion mining that identifies emotions or opinions contained in texts, and Text Summarization which summarize the main contents from one document or several documents have been actively studied. Especially, the text summarization technique is actively applied in the business through the news summary service, the privacy policy summary service, ect. In addition, much research has been done in academia in accordance with the extraction approach which provides the main elements of the document selectively and the abstraction approach which extracts the elements of the document and composes new sentences by combining them. However, the technique of evaluating the quality of automatically summarized documents has not made much progress compared to the technique of automatic text summarization. Most of existing studies dealing with the quality evaluation of summarization were carried out manual summarization of document, using them as reference documents, and measuring the similarity between the automatic summary and reference document. Specifically, automatic summarization is performed through various techniques from full text, and comparison with reference document, which is an ideal summary document, is performed for measuring the quality of automatic summarization. Reference documents are provided in two major ways, the most common way is manual summarization, in which a person creates an ideal summary by hand. Since this method requires human intervention in the process of preparing the summary, it takes a lot of time and cost to write the summary, and there is a limitation that the evaluation result may be different depending on the subject of the summarizer. Therefore, in order to overcome these limitations, attempts have been made to measure the quality of summary documents without human intervention. On the other hand, as a representative attempt to overcome these limitations, a method has been recently devised to reduce the size of the full text and to measure the similarity of the reduced full text and the automatic summary. In this method, the more frequent term in the full text appears in the summary, the better the quality of the summary. However, since summarization essentially means minimizing a lot of content while minimizing content omissions, it is unreasonable to say that a "good summary" based on only frequency always means a "good summary" in its essential meaning. In order to overcome the limitations of this previous study of summarization evaluation, this study proposes an automatic quality evaluation for text summarization method based on the essential meaning of summarization. Specifically, the concept of succinctness is defined as an element indicating how few duplicated contents among the sentences of the summary, and completeness is defined as an element that indicating how few of the contents are not included in the summary. In this paper, we propose a method for automatic quality evaluation of text summarization based on the concepts of succinctness and completeness. In order to evaluate the practical applicability of the proposed methodology, 29,671 sentences were extracted from TripAdvisor 's hotel reviews, summarized the reviews by each hotel and presented the results of the experiments conducted on evaluation of the quality of summaries in accordance to the proposed methodology. It also provides a way to integrate the completeness and succinctness in the trade-off relationship into the F-Score, and propose a method to perform the optimal summarization by changing the threshold of the sentence similarity.

• Journal of Intelligence and Information Systems
• /
• v.27 no.3
• /
• pp.95-112
• /
• 2021

In university education, the choice of major class plays an important role in students' careers. However, in line with the changes in the industry, the fields of major subjects by department are diversifying and increasing in number in university education. As a result, students have difficulty to choose and take classes according to their career paths. In general, students choose classes based on experiences such as choices of peers or advice from seniors. This has the advantage of being able to take into account the general situation, but it does not reflect individual tendencies and considerations of existing courses, and has a problem that leads to information inequality that is shared only among specific students. In addition, as non-face-to-face classes have recently been conducted and exchanges between students have decreased, even experience-based decisions have not been made as well. Therefore, this study proposes a recommendation system model that can recommend college major classes suitable for individual characteristics based on data rather than experience. The recommendation system recommends information and content (music, movies, books, images, etc.) that a specific user may be interested in. It is already widely used in services where it is important to consider individual tendencies such as YouTube and Facebook, and you can experience it familiarly in providing personalized services in content services such as over-the-top media services (OTT). Classes are also a kind of content consumption in terms of selecting classes suitable for individuals from a set content list. However, unlike other content consumption, it is characterized by a large influence of selection results. For example, in the case of music and movies, it is usually consumed once and the time required to consume content is short. Therefore, the importance of each item is relatively low, and there is no deep concern in selecting. Major classes usually have a long consumption time because they have to be taken for one semester, and each item has a high importance and requires greater caution in choice because it affects many things such as career and graduation requirements depending on the composition of the selected classes. Depending on the unique characteristics of these major classes, the recommendation system in the education field supports decision-making that reflects individual characteristics that are meaningful and cannot be reflected in experience-based decision-making, even though it has a relatively small number of item ranges. This study aims to realize personalized education and enhance students' educational satisfaction by presenting a recommendation model for university major class. In the model study, class history data of undergraduate students at University from 2015 to 2017 were used, and students and their major names were used as metadata. The class history data is implicit feedback data that only indicates whether content is consumed, not reflecting preferences for classes. Therefore, when we derive embedding vectors that characterize students and classes, their expressive power is low. With these issues in mind, this study proposes a Net-NeuMF model that generates vectors of students, classes through network analysis and utilizes them as input values of the model. The model was based on the structure of NeuMF using one-hot vectors, a representative model using data with implicit feedback. The input vectors of the model are generated to represent the characteristic of students and classes through network analysis. To generate a vector representing a student, each student is set to a node and the edge is designed to connect with a weight if the two students take the same class. Similarly, to generate a vector representing the class, each class was set as a node, and the edge connected if any students had taken the classes in common. Thus, we utilize Node2Vec, a representation learning methodology that quantifies the characteristics of each node. For the evaluation of the model, we used four indicators that are mainly utilized by recommendation systems, and experiments were conducted on three different dimensions to analyze the impact of embedding dimensions on the model. The results show better performance on evaluation metrics regardless of dimension than when using one-hot vectors in existing NeuMF structures. Thus, this work contributes to a network of students (users) and classes (items) to increase expressiveness over existing one-hot embeddings, to match the characteristics of each structure that constitutes the model, and to show better performance on various kinds of evaluation metrics compared to existing methodologies.

• The Journal of Korean Society for Radiation Therapy
• /
• v.27 no.1
• /
• pp.31-43
• /
• 2015

Purpose : Stereotactic body radiation therapy (SBRT) has proved its efficacy in several patient populations with primary and metastatic limited tumors. Because SBRT prescription is high dose level than Conventional radiation therapy. SBRT plan is necessary for effective Organ at risk (OAR) protection and sufficient Planning target volume (PTV) dose coverage. In particular, multi-target cases may result excessive doses to OAR and hot spot due to dose overlap. This study evaluate usefulness of Volumetric modulated arc therapy (VMAT) in dosimetric and technical considerations using Flattening filter free (FFF) beam. Materials and Methods : The treatment plans for five patients, being treated on TrueBeam STx(Varian$^{TM}$, USA) with VMAT using 10MV FFF beam and Standard conformal radiotherapy (CRT) using 15MV Flattening filter (FF) beam. PTV, liver, duodenum, bowel, spinal cord, esophagus, stomach dose were evaluated using the dose volume histogram(DVH). Conformity index(CI), homogeneity index(HI), Paddick's index(PCI) for the PTV was assessed. Total Monitor unit (MU) and beam on time was assessed. Results : Average value of CI, HI and PCI for PTV was $1.381{\pm}0.028$, $1.096{\pm}0.016$, $0.944{\pm}0.473$ in VMAT and $1.381{\pm}0.042$, $1.136{\pm}0.042$, $1.534{\pm}0.465$ in CRT respectively. OAR dose in CRT plans evaluated 1.8 times higher than VMAT. Total MU in VMAT evaluated 1.3 times increase than CRT. Average beam on time was 6.8 minute in VMAT and 21.3 minute in CRT. Conclusion : VMAT for SBRT in multi-target liver cancer using FFF beam is effective treatment techniqe in dosimetric and technical considerations. VMAT decrease intra-fraction error due to treatment time shortening using high dose rate of FFF beam.

• Journal of Nutrition and Health
• /
• v.43 no.4
• /
• pp.404-412
• /
• 2010

This paper specifically discusses the risk assessment on the pesticide residues in vegetables collected from traditional markets, big marts and departments in the southern part of Seoul. Vegetable samples were 6,583 cases from January to December in 2009. Monte-Carlo simulation was used to calculate the uncertainty for the risk index using pesticide residues, average dietary intake for vegetables and acceptable daily intake. Deterministic risk indexes were 7.33% of diethofencarb, 5.13% of indoxacarb, 3.96% of EPN, 3.92% of diniconazole and 3.09% of chlorothalonil, respectively. And other pesticides were below 3%. Distributions of risk indexes obtained by the Monte-Carlo simulations were similar to the deterministic values, even though the confidence intervals for 95% were very wide. We confirmed that health risks caused by eating vegetables exceeded maximum residue limits of pesticide are very low and the population is generally safe, judging from the risk indexes located between 0.07 to 9.49%.

• Journal of Korea Spatial Information System Society
• /
• v.9 no.1
• /
• pp.105-115
• /
• 2007

Data mining extracts interesting knowledge from a large database. Among numerous data mining techniques, research work is primarily concentrated on clustering and association rules. The clustering technique of the active research topics mainly deals with analyzing spatial and attribute data. And, the technique of association rules deals with identifying frequent patterns. There was an advanced apriori algorithm using an existing bit-clustering algorithm. In an effort to identify an alternative algorithm to improve apriori, we investigated FP-Growth and discussed the possibility of adopting bit-clustering as the alternative method to solve the problems with FP-Growth. FP-Growth using bit-clustering demonstrated better performance than the existing method. We used chess data in our experiments. Chess data were used in the pattern mining evaluation. We made a creation of FP-Tree with different minimum support values. In the case of high minimum support values, similar results that the existing techniques demonstrated were obtained. In other cases, however, the performance of the technique proposed in this paper showed better results in comparison with the existing technique. As a result, the technique proposed in this paper was considered to lead to higher performance. In addition, the method to apply bit-clustering to GML data was proposed.

• Journal of Korea Water Resources Association
• /
• v.46 no.6
• /
• pp.597-611
• /
• 2013

This paper presents the methodology for construction of time-area curve via the width function and thereby rational estimation of time of concentration and storage coefficient of Clark model within the framework of method of moments. To this end time-area curve is built by rescaling the grid-based width function under the assumption of pure translation and then the analytical expressions for two parameters of Clark model are proposed in terms of method of moments. The methodology in this study based on the analytical expressions mentioned before is compared with both (1) the traditional optimization method of Clark model provided by HEC-1 in which the symmetric time-area curve is used and the difference between observed and simulated hydrographs is minimized (2) and the same optimization method but replacing time-area curve with rescaled width function in respect of peak discharge and time to peak of simulated direct runoff hydrographs and their efficiency coefficient relative to the observed ones. The following points are worth of emphasizing: (1) The optimization method by HEC-1 with rescaled width function among others results in the parameters well reflecting the observed runoff hydrograph with respect to peak discharge coordinates and coefficient of efficiency; (2) For the better application of Clark model it is recommended to use the time-area curve capable of accounting for irregular drainage structure of a river basin such as rescaled width function instead of symmetric time-area curve by HEC-1; (3) Moment-based methodology with rescaled width function developed in this study also gives rise to satisfactory simulation results in terms of peak discharge coordinates and coefficient of efficiency. Especially the mean velocities estimated from this method, characterizing the translation effect of time-area curve, are well consistent with the field surveying results for the points of interest in this study; (4) It is confirmed that the moment-based methodology could be an effective tool for quantitative assessment of translation and storage effects of natural river basin; (5) The runoff hydrographs simulated by the moment-based methodology tend to be more right skewed relative to the observed ones and have lower peaks. It is inferred that this is due to consideration of only one mean velocity in the parameter estimation. Further research is required to combine the hydrodynamic heterogeneity between hillslope and channel network into the construction of time-area curve.

• The Journal of Korean Society for Radiation Therapy
• /
• v.29 no.2
• /
• pp.83-91
• /
• 2017

Objectives: The purpose of this study was to evaluate the error of the leaf position accuracy of the MLC due to the gravity effect according to the gantry angle by using picket fence test using EPID and GafChromic EBT3 film. Materials and Methods: A 5 cm solid phantom was placed on the table and the SAD was set to 100 cm. The EBT3 film was placed exactly over the solid phantom and covered a 1.5 cm solid phantom and the picket fence test was performed. The EPID was measured under the same conditions as the EBT3 film at SID 100 cm. The gantry angles were measured at $0^{\circ}$, $90^{\circ}$, $180^{\circ}$ and $270^{\circ}$ in order to evaluate the position of the MLC according to the gantry angle. For the geometric evaluation of the MLC, the leaf position accuracy of the MLC was analyzed using the analysis program. Results: In case of EPID, when the gantry angle was changed to $0^{\circ}$, $90^{\circ}$, $180^{\circ}$, $270^{\circ}$, the difference of the position errors of the leaves was 0.18 mm, 0.31 mm, 0.20 mm, 0.26 mm on the average and the maximum values of the errors were respectively 0.44 mm, 0.54 mm, 0.34 mm, 0.44 mm. In case of EBT3 film, when the gantry angle was changed to $0^{\circ}$, $90^{\circ}$, $180^{\circ}$, $270^{\circ}$, the difference of the position errors of the leaves was 0.19 mm, 0.21 mm, 0.19 mm, 0.31 mm on the average and the maximum values of the errors were respectively 0.35 mm, 0.45 mm, 0.36 mm, 0.48 mm. Conclusion: In this study, we analyzed the position error of the leaf of the MLC according to the gantry angle, and confirmed the position error of the leaf by gravity effect. As a result of comparing the leaf position accuracy using EPID and EBT3 film according to the variation of gantry angle, a larger error occurred in the error analysis method using EPID than that of EBT3 film. Therefore, in the case of IMRT based on MLC, as well as verification of accurate dosimetry should be conducted, it is considered that the quality control and verification for the precise operation of the MLC will be needed. and it is necessary to compare and verify the method of analysis.

• Korean Journal of Poultry Science
• /
• v.44 no.4
• /
• pp.253-258
• /
• 2017

This study was aimed at evaluating the additivity of crude protein digestibility in mixed diets of corn and soybean meal (SBM), and comparing direct and indirect methods for evaluating crude protein (CP) digestibility. Totally, five hundred and twenty-five 18-day-old broiler chickens were grouped into 7 blocks based on body weight, and randomly allocated to 6 treatment groups in a randomized complete block design. The basal diet, diet 3, was corn-SBM-based, containing 65% corn and 28% SBM. Diets 1 and 5 contained corn and SBM, respectively, as the sole CP source. To use the difference method, 2 diets, diets 2 and 4, were prepared by mixing corn and SBM at the expense of the basal diet, respectively. Diet 2 contained 79% corn and 14% SBM, and diet 4 contained 32.5% corn and 34% SBM. To evaluate the additivity of digestibility values, the difference between measured values for the mixed diets (diets 2, 3, and 4) and predicted values calculated using the measured values for diets 1 and 5 was examined. The apparent (AID) and standardized (SID) ileal digestibility of CP in SBM differed between the direct and indirect methods; however, corresponding digestibility did not differ for corn. Additionally, the predicted and measured digestibility of both AID and SID differed in diets 2 and 3, implying that the digestibility values obtained from diets 1 and 5 were not additive for mixed diets. In conclusion, this study showed that digestibility evaluated by direct and indirect methods depends on the ingredients having different CP concentrations, and this finding may be considered to improve the accuracy of feed formulation for broiler chickens.

• Journal of Korean Society of Environmental Engineers
• /
• v.37 no.8
• /
• pp.480-491
• /
• 2015

In this study, methods using LDC (Load Duration Curve) and watershed model were suggested to develope management targets and evaluate target achievement for non-point source pollution management considering watershed and runoff characteristics and possibility for achievement of target. These methods were applied for Saemangeum watershed which was designated as nonpoint source pollution management area recently. Flow duration interval of 5 to 40% was selected as flow range for management considering runoff characteristics and TP was selected as indicator for management. Management targets were developed based on scenarios for non-point source pollutant reduction of management priority areas using LDC method and HSPF model which was calibrated using 4 years data (2009~2012). In the scenario of LID, road sweeping and 50% reduction in CSOs and untreated sewage at Jeonju A20 and 30% reduction in fertilizer and 50% in livestock NPS at Mankyung C03, Dongjin A14 and KobuA14, management targets for Mangyung bridge, Dongjin bridge, Jeonju stream and Gunpo bridge were developed as TP 0.38, 0.18, 0.64 and 0.16 mg/L respectively. When TP loads at the target stations were assumed to have been reduced by a certain percentage (10%), management targets for those target stations were developed as TP 0.35, 0.17, 0.60 and 0.15 mg/L respectively. The result of this study is expected to be used as reference material for management master plan, implementation plan and implementation assessment for non-point source management area.

• The Journal of Korean Society for Radiation Therapy
• /
• v.20 no.1
• /
• pp.37-43
• /
• 2008

Purpose: To evaluate dosimetry results of three different techniques for whole breast irradiation after conservative surgery of large pendulous breast patient. Materials and Methods: Planning computed tomography (CT) scans for three techniques were performed on a GE Hi-speed advantage CT scanner in the supine (SP), supine with breast supporting Device (SD) and prone position on a custom prone mattress (PP). Computed tomography images were acquired at 5 mm thickness. The clinical target volumes (CTV), ipsilateral lung and heart were delineated to evaluate the dose statistic, and all techniques were planned with the tangential photon beams (Pinnacle$^3$, Philips Medical System, USA). The prescribed dose was 50 Gy delivered in 25 fractions. To evaluate the dose coverage for CTV, we analysed percent volume of CTV receiving minimum of 95%, 100%, 105%, and 110% of prescription dose ($V_{95}$, $V_{100}$, $V_{105}$, and $V_{110}$) and minimal dose covering 95% ($D_{95}$) of CTV. The dosimetric comparison for heart and ipsilateral lung was analysed using the minimal dose covering 5% of each organs ($D_5$) and the volume that received >18 Gy for the heart and >20 Gy for the ipsilateral lung. Results: Target volume coverage ($V_{95}$ and $V_{100}$) was not significantly different for all technique. The V105 was lower for PP (1.2% vs. 4.4% for SP, 11.1% for SD). Minimal dose covering 95% ($D_{95}$) of target was 47.5 Gy, 47.7 Gy and 48 Gy for SP, SD and PP. The volume of ipsilateral lung received >20 Gy was 21.7%, 11.6% and 4.9% for SP, SD and PP. The volume of heart received >18 Gy was 17.0%, 16.1% and 9.8% for SP, SD and PP. Conclusion: Prone positioning of patient for large pendulous breast irradiation enables improving dose uniformity with minimal heart and lung doses.