• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.67 no.4
• /
• pp.265-273
• /
• 2022

Waxy corn (Zea mays L.), which contains homozygous mutant alleles for the waxy1 (wx1) gene, is widely consumed as a snack food in Asia. This study evaluated sixteen agronomic characteristics of inbred Korean waxy corn lines to aid development of high-quality waxy corn cultivars. The plant materials studied were 177 inbred waxy corn lines developed by the National Institute of Crop Science, Rural Development Administration, Republic of Korea. For the tested lines, days to tasseling and silking averaged 77.69±2.22 days (with a range of 56-97 days), and 81.12±7.56 days (66-99 days), respectively. Plant length ranged from 88 to 237 cm (averaged 164.88±22.67 cm), ear length averaged 11.75±2.52 cm (5.0-18.5 cm), and ear width averaged 2.94±0.68 cm (1.4-4.5 cm). The number of rows on each ear of corn averaged 12.22±2.22 (7-32 rows) and the kernel number averaged 24.30±4.22 (9-37 kernels) per row. The crude protein content was 12.05±1.53% (8.90-21.80%) and total starch content was 69.27±5.74% (49.5-83.9%). Principal component analysis revealed that ear width, grain length, ear length, days to tasseling, days to silking, percentage of ear setting height, and total starch are features that allow distinction between the 177 waxy inbred corn lines. Hierarchical cluster analysis identified twelve waxy inbred lines that produce tall plants and have a short silking period. These lines may improve yield among quickly growing corn varieties.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.249-263
• /
• 2023

Collaborative Filtering, a representative recommendation system methodology, consists of two approaches: neighbor methods and latent factor models. Among these, the latent factor model using matrix factorization decomposes the user-item interaction matrix into two lower-dimensional rectangular matrices, predicting the item's rating through the product of these matrices. Due to the factor vectors inferred from rating patterns capturing user and item characteristics, this method is superior in scalability, accuracy, and flexibility compared to neighbor-based methods. However, it has a fundamental drawback: the need to reflect the diversity of preferences of different individuals for items with no ratings. This limitation leads to repetitive and inaccurate recommendations. The Adaptive Deep Latent Factor Model (ADLFM) was developed to address this issue. This model adaptively learns the preferences for each item by using the item description, which provides a detailed summary and explanation of the item. ADLFM takes in item description as input, calculates latent vectors of the user and item, and presents a method that can reflect personal diversity using an attention score. However, due to the requirement of a dataset that includes item descriptions, the domain that can apply ADLFM is limited, resulting in generalization limitations. This study proposes a Generalized Adaptive Deep Latent Factor Recommendation Model, G-ADLFRM, to improve the limitations of ADLFM. Firstly, we use item ID, commonly used in recommendation systems, as input instead of the item description. Additionally, we apply improved deep learning model structures such as Self-Attention, Multi-head Attention, and Multi-Conv1D. We conducted experiments on various datasets with input and model structure changes. The results showed that when only the input was changed, MAE increased slightly compared to ADLFM due to accompanying information loss, resulting in decreased recommendation performance. However, the average learning speed per epoch significantly improved as the amount of information to be processed decreased. When both the input and the model structure were changed, the best-performing Multi-Conv1d structure showed similar performance to ADLFM, sufficiently counteracting the information loss caused by the input change. We conclude that G-ADLFRM is a new, lightweight, and generalizable model that maintains the performance of the existing ADLFM while enabling fast learning and inference.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.307-325
• /
• 2023

Due to the prolonged COVID-19 pandemic, the frequency of people who are tired of living indoors visiting nearby mountains and national parks to relieve depression and lethargy has exploded. There is a place where thousands of people who came out of nature stop walking and breathe and rest, that is the mineral spring. Even in mountains or national parks, there are about 600 mineral springs that can be found occasionally in neighboring parks or trails in the metropolitan area. However, due to irregular and manual water quality tests, people drink mineral water without knowing the test results in real time. Therefore, in this study, we intend to develop a model that can predict the quality of the spring water in real time by exploring the factors affecting the quality of the spring water and collecting data scattered in various places. After limiting the regions to Seoul and Gyeonggi-do due to the limitations of data collection, we obtained data on water quality tests from 2015 to 2020 for about 300 mineral springs in 18 cities where data management is well performed. A total of 10 factors were finally selected after two rounds of review among various factors that are considered to affect the suitability of the mineral spring water quality. Using AutoML, an automated machine learning technology that has recently been attracting attention, we derived the top 5 models based on prediction performance among about 20 machine learning methods. Among them, the catboost model has the highest performance with a prediction classification accuracy of 75.26%. In addition, as a result of examining the absolute influence of the variables used in the analysis through the SHAP method on the prediction, the most important factor was whether or not a water quality test was judged nonconforming in the previous water quality test. It was confirmed that the temperature on the day of the inspection and the altitude of the mineral spring had an influence on whether the water quality was unsuitable.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.121-142
• /
• 2023

As the number and weight of imported food are steadily increasing, safety management of imported food to prevent food safety accidents is becoming more important. The Ministry of Food and Drug Safety conducts on-site inspections of foreign food facilities before customs clearance as well as import inspection at the customs clearance stage. However, a data-based safety management plan for imported food is needed due to time, cost, and limited resources. In this study, we tried to increase the efficiency of the on-site inspection by preparing a machine learning prediction model that pre-selects the companies that are expected to fail before the on-site inspection. Basic information of 303,272 foreign food facilities and processing businesses collected in the Integrated Food Safety Information Network and 1,689 cases of on-site inspection information data collected from 2019 to April 2022 were collected. After preprocessing the data of foreign food facilities, only the data subject to on-site inspection were extracted using the foreign food facility_code. As a result, it consisted of a total of 1,689 data and 103 variables. For 103 variables, variables that were '0' were removed based on the Theil-U index, and after reducing by applying Multiple Correspondence Analysis, 49 characteristic variables were finally derived. We build eight different models and perform hyperparameter tuning through 5-fold cross validation. Then, the performance of the generated models are evaluated. The research purpose of selecting companies subject to on-site inspection is to maximize the recall, which is the probability of judging nonconforming companies as nonconforming. As a result of applying various algorithms of machine learning, the Random Forest model with the highest Recall_macro, AUROC, Average PR, F1-score, and Balanced Accuracy was evaluated as the best model. Finally, we apply Kernal SHAP (SHapley Additive exPlanations) to present the selection reason for nonconforming facilities of individual instances, and discuss applicability to the on-site inspection facility selection system. Based on the results of this study, it is expected that it will contribute to the efficient operation of limited resources such as manpower and budget by establishing an imported food management system through a data-based scientific risk management model.

• Journal of Intelligence and Information Systems
• /
• v.29 no.1
• /
• pp.143-174
• /
• 2023

With the development of deep learning technologies, Artificial Intelligence powered Optical Character Recognition (AI-OCR) has evolved to read multiple languages from various forms of images accurately. For the financial industry, where a large number of diverse documents are processed through manpower, the potential for using AI-OCR is great. In this study, we present a configuration and a design of an AI-OCR modality for use in the financial industry and discuss the platform construction with application cases. Since the use of financial domain data is prohibited under the Personal Information Protection Act, we developed a deep learning-based data generation approach and used it to train the AI-OCR models. The AI-OCR models are trained for image preprocessing, text recognition, and language processing and are configured as a microservice architected platform to process a broad variety of documents. We have demonstrated the AI-OCR platform by applying it to financial domain tasks of document sorting, document verification, and typing assistance The demonstrations confirm the increasing work efficiency and conveniences.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.67 no.4
• /
• pp.342-361
• /
• 2022

In order to evaluate the effect of nitrogen application levels on yield and quality of rice varieties, a field experiment was conducted at National Institute of Crop Science of Korea from 2018 to 2020. Five levels (0, 3, 5, 7, and 9 kg/10a) of nitrogen fertilizer were treated to 21 Korean rice varieties. Yield, yield component, appearance quality, and protein content in rice were analyzed. The average head rice yield for 3 years decreased by 28%, 22%, 11%, and 8%, respectively, when cultivated with 0, 3, 5, and 7 kg/10a nitrogen application compared to cultivation with a standard nitrogen application amount, 9 kg/10a. The number of panicles per hill increased as the amount of nitrogen application increased, but there was no significant change in the number of grains per panicle and 1000-grains weight, and the number of panicles per hill showed relatively small annual variation compared to other yield components. There was no significant difference in the head rice ratio according to the nitrogen application amount, the broken rice ratio slightly decreased, and the floury rice ratio increased. The protein content of rice decreased with increasing nitrogen application in 2018 and 2019, and was the lowest at 7 kg/10a of nitrogen application, and showed a tendency to increase again at 9 kg/10a. In the case of 2020, as the amount of nitrogen application increased, the protein content showed a tendency to continuously increase. In terms of varieties, 13 varieties, including Chilbo, seemed to be capable of low-nitrogen cultivation because loss of the head rice yield was less and the protein content could be lowered to 6% or less according to 7 kg/10a nitrogen application.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.38 no.3
• /
• pp.264-274
• /
• 1993

This study was carried out to assess the agronomic characters and physicochemical properties of floury and chalky-endosperm mutant lines induced by chemical mutagen treatment to rice varieties, Hwacheongbyeo and IR24. Linkage analysis of a floury-endosperm gene was carried out using linkage testers. The grain size of brown rice of the mutants was smaller than that of the original varieties. The l, 000-grain and 1$\ell$ weight were lighter in the mutants compared with those in the original varieties. The compound starch granules in the endosperm cell of the mutants showed a loosely-packed crystalline structure. Amylose contents in mutants ranged from 16.9 to 28.5%. Crude protein contents of the mutants were not significantly different from the original rice variety, Hwacheongbyeo, but white core mutant(line 47106) derived from IR24 showed higher protein(l1.32%) compared with IR24(8.30%). The mutants showed slightly harder gel characteristics, and much lower viscosity in Amylograph than original varieties. Steamed rice-cakes from mutant lines showed greater volume than those from original varieties. During the process of alcohol fermentation, Brix in the mutants(especially floury mutants) decreased faster and the alcohol production after 10-day fermentation was much greater in the mutants than in the original varieties. Three different gene loci for floury endosperm characteristics were identified from the allelism test among mutant lines, and the genes were tentatively symbolized as flo-a, flo-b and flo-c, respectively. A floury gene, flo-a, was linked with lg(liguleless) gene in the linkage group N, with R.V. 5.76$\pm$1.72%.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.14
• /
• pp.159-164
• /
• 1973

This study was carried out in the central part of Korea for three years from 1970 to 1972. Its aim was to investigate the variations of growth and yield of rice varieties of different cultivating methods (transplanting culture, direct sowing culture on irrigated paddy field, and direct sowing culture on non-irrigated paddy field. The results could be summarized as follows; 1. The number of days from sowing to heading by direct sowing culture on irrigated paddy field was remarkably shorter than transplanting culture. Direct sowing culture on non-irrigated paddy field was longer than transplanting culture. 2. The number of panicles showed direct sowing culture on irrigated paddy field was first, direct sowing culture on non-irrigated paddy field was second, and third transplanting culture. But the number of spikelets per panicle was first on the transplanting culture, second on the direct sowing culture on non-irrigated paddy field and third on the direct sowing culture on irrigated paddy field. 3. The weight of the brown rice showed that direct sowing culture on irrigated paddy field produced the highest yield among the cultivating method in all varieties. The transplanting culture increased more than the direct sowing culture on non-irrigated paddy field. The highly productive varieties were Sadominory, Akibale, Suwon 213-1 when the direct sowing culture on irrigated paddy field was used. 4. The correlation between the brown rice yield and the yield components were as follows; in each cultivating method, it showed highly significant positive correlation between the brown rice yield and the 1, 000 grain weight of hulled rice, and between the brown rice yield and the number of the spiklets per panicle in the transplanting culture, in the direct sowing culture On irrigated paddy field between the brown rice yield and the number of spikelets per panicle and between the brown rice yield and the number of panicles, in direct sowing culture on non-irrigated paddy field between the brown rice yield and number of panicles. 5. It showed highly significant negative correlation between the brown rice yield and the number of days from sowing to heading in the direct sowing culture on irrigated and non-irrigated paddy field, but it was showed very low negative correlation in the transplant culture.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.26 no.2
• /
• pp.15-19
• /
• 2022

Purpose [⁶⁸Ga]PSMA-11 is needed the high reproducibility, excellent radiochemical yield and purity. In term of radiation safety, the radiation exposure of operator for its production also should be considered. In this work, we performed a comparative study for the fully automated synthesis of [⁶⁸Ga]PSMA-11 between non-cassette type and cassette type. Materials and Methods Two different type of modules (TRACERlab FX N pro for non-cassette type and BIKBox for cassette type) were used for the automated production of [⁶⁸Ga]PSMA-11. According to the previously identified elution profile, Only 2.5 ml with high radioactivity was used for the reaction. After adjusting the pH of the reaction solution with HEPES buffer solution, the precursor was added and reacted with at 95 ℃ for 15 minutes. The reaction mixture was separated and purified using a C18 light cartridge. The product was eluted with 50% EtOH/saline solution and diluted with saline. It was completed by sterilizing filter. In the non-cassette type, the aforementioned process must be prepared directly. However, in the cassette method, synthesis was possible simply by installing a kit that was already completed. Results Both total [⁶⁸Ga]PSMA-11 production time were 25±3(non-cassette type) and 23±3 minutes(cassette type). The radiochemical yield of the non-cassette type(65.5±5.7%) was higher than that of the cassette type(61.6±4.8%) after sterilization filter. The non-cassette type took about 120 minutes of preparation time before synthesis due to washing of synthesizer and reagent preparation. However, since the cassette type does not require washing and reagent preparation, it took about 20 minutes to prepare before synthesis. Both type of synthesizer had a radiochemical high purity(>99%). Conclusion The non-cassette type production of [⁶⁸Ga]PSMA-11 showed higher radiochemical yield and lower cost than the cassette type. However, The cassette type has an advantage in terms of preparation time, convenience, and equipment maintenance.

• Journal of Nutrition and Health
• /
• v.56 no.1
• /
• pp.35-53
• /
• 2023

Purpose: The purpose of the first part of this study was to evaluate the validity of the physical activity classification table for youth (Youth-PACT). The second part of this study was aimed at comparing the estimated energy requirement (EER) with the total energy expenditure (TEE) and evaluating the physical activity patterns of Korean children and adolescents. Methods: The subjects of the first part of the study were 17 children aged 10 to 12 years, and their total energy expenditure (TEE_DLW) was measured using the double labeled water (DLW) method. A total of 166 children and adolescents aged 6-18 years participated in the second part of this study. Their resting energy expenditure (REE) was measured using indirect calorimetry and the TEE_Youth-PACT and physical activity level were calculated by applying the Youth-PACT to the physical activity diary prepared by the subjects. Results: In the first part of this study, there were no significant differences between the TEE_DLW and the TEE_Youth-PACT. The TEE_Youth-PACT accurately predicted TEE_DLW in 37.5% of the subjects. In the second part of the study, the rates at which EER accurately predicted TEE _YouthPACT and overestimated TEE _Youth-PACT were 29.6% and 47.3%, respectively. The time spent based on intensity of physical activity and the physical activity categories which were obtained using Youth-PACT showed different patterns according to sex and age group. Age showed significant positive correlations with REE, TEE, and the time spent in sedentary behavior, but age was significantly negatively correlated with REE/body weight, TEE/body weight, and the time spent in low-intensity and high-intensity activities. Conclusion: The results of this study showed that the Youth-PACT can be used to evaluate the TEE and PAL of children and adolescents. However, further studies are needed to validate the TEE_Youth-PACT and to set the EER for children and adolescents.