• Journal of Intelligence and Information Systems
• /
• v.26 no.4
• /
• pp.1-25
• /
• 2020

Sentiment Analysis (SA) is a Natural Language Processing (NLP) task that analyzes the sentiments consumers or the public feel about an arbitrary object from written texts. Furthermore, Aspect-Based Sentiment Analysis (ABSA) is a fine-grained analysis of the sentiments towards each aspect of an object. Since having a more practical value in terms of business, ABSA is drawing attention from both academic and industrial organizations. When there is a review that says "The restaurant is expensive but the food is really fantastic", for example, the general SA evaluates the overall sentiment towards the 'restaurant' as 'positive', while ABSA identifies the restaurant's aspect 'price' as 'negative' and 'food' aspect as 'positive'. Thus, ABSA enables a more specific and effective marketing strategy. In order to perform ABSA, it is necessary to identify what are the aspect terms or aspect categories included in the text, and judge the sentiments towards them. Accordingly, there exist four main areas in ABSA; aspect term extraction, aspect category detection, Aspect Term Sentiment Classification (ATSC), and Aspect Category Sentiment Classification (ACSC). It is usually conducted by extracting aspect terms and then performing ATSC to analyze sentiments for the given aspect terms, or by extracting aspect categories and then performing ACSC to analyze sentiments for the given aspect category. Here, an aspect category is expressed in one or more aspect terms, or indirectly inferred by other words. In the preceding example sentence, 'price' and 'food' are both aspect categories, and the aspect category 'food' is expressed by the aspect term 'food' included in the review. If the review sentence includes 'pasta', 'steak', or 'grilled chicken special', these can all be aspect terms for the aspect category 'food'. As such, an aspect category referred to by one or more specific aspect terms is called an explicit aspect. On the other hand, the aspect category like 'price', which does not have any specific aspect terms but can be indirectly guessed with an emotional word 'expensive,' is called an implicit aspect. So far, the 'aspect category' has been used to avoid confusion about 'aspect term'. From now on, we will consider 'aspect category' and 'aspect' as the same concept and use the word 'aspect' more for convenience. And one thing to note is that ATSC analyzes the sentiment towards given aspect terms, so it deals only with explicit aspects, and ACSC treats not only explicit aspects but also implicit aspects. This study seeks to find answers to the following issues ignored in the previous studies when applying the BERT pre-trained language model to ACSC and derives superior ACSC models. First, is it more effective to reflect the output vector of tokens for aspect categories than to use only the final output vector of [CLS] token as a classification vector? Second, is there any performance difference between QA (Question Answering) and NLI (Natural Language Inference) types in the sentence-pair configuration of input data? Third, is there any performance difference according to the order of sentence including aspect category in the QA or NLI type sentence-pair configuration of input data? To achieve these research objectives, we implemented 12 ACSC models and conducted experiments on 4 English benchmark datasets. As a result, ACSC models that provide performance beyond the existing studies without expanding the training dataset were derived. In addition, it was found that it is more effective to reflect the output vector of the aspect category token than to use only the output vector for the [CLS] token as a classification vector. It was also found that QA type input generally provides better performance than NLI, and the order of the sentence with the aspect category in QA type is irrelevant with performance. There may be some differences depending on the characteristics of the dataset, but when using NLI type sentence-pair input, placing the sentence containing the aspect category second seems to provide better performance. The new methodology for designing the ACSC model used in this study could be similarly applied to other studies such as ATSC.

• Journal of Intelligence and Information Systems
• /
• v.27 no.4
• /
• pp.73-95
• /
• 2021

This study uses Node2vec graph embedding method and Light GBM link prediction to explore undeveloped export candidate countries in Korea's food and beverage industry. Node2vec is the method that improves the limit of the structural equivalence representation of the network, which is known to be relatively weak compared to the existing link prediction method based on the number of common neighbors of the network. Therefore, the method is known to show excellent performance in both community detection and structural equivalence of the network. The vector value obtained by embedding the network in this way operates under the condition of a constant length from an arbitrarily designated starting point node. Therefore, it has the advantage that it is easy to apply the sequence of nodes as an input value to the model for downstream tasks such as Logistic Regression, Support Vector Machine, and Random Forest. Based on these features of the Node2vec graph embedding method, this study applied the above method to the international trade information of the Korean food and beverage industry. Through this, we intend to contribute to creating the effect of extensive margin diversification in Korea in the global value chain relationship of the industry. The optimal predictive model derived from the results of this study recorded a precision of 0.95 and a recall of 0.79, and an F1 score of 0.86, showing excellent performance. This performance was shown to be superior to that of the binary classifier based on Logistic Regression set as the baseline model. In the baseline model, a precision of 0.95 and a recall of 0.73 were recorded, and an F1 score of 0.83 was recorded. In addition, the light GBM-based optimal prediction model derived from this study showed superior performance than the link prediction model of previous studies, which is set as a benchmarking model in this study. The predictive model of the previous study recorded only a recall rate of 0.75, but the proposed model of this study showed better performance which recall rate is 0.79. The difference in the performance of the prediction results between benchmarking model and this study model is due to the model learning strategy. In this study, groups were classified by the trade value scale, and prediction models were trained differently for these groups. Specific methods are (1) a method of randomly masking and learning a model for all trades without setting specific conditions for trade value, (2) arbitrarily masking a part of the trades with an average trade value or higher and using the model method, and (3) a method of arbitrarily masking some of the trades with the top 25% or higher trade value and learning the model. As a result of the experiment, it was confirmed that the performance of the model trained by randomly masking some of the trades with the above-average trade value in this method was the best and appeared stably. It was found that most of the results of potential export candidates for Korea derived through the above model appeared appropriate through additional investigation. Combining the above, this study could suggest the practical utility of the link prediction method applying Node2vec and Light GBM. In addition, useful implications could be derived for weight update strategies that can perform better link prediction while training the model. On the other hand, this study also has policy utility because it is applied to trade transactions that have not been performed much in the research related to link prediction based on graph embedding. The results of this study support a rapid response to changes in the global value chain such as the recent US-China trade conflict or Japan's export regulations, and I think that it has sufficient usefulness as a tool for policy decision-making.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.23 no.2
• /
• pp.51-58
• /
• 2019

Purpose In in-vitro laboratories of nuclear medicine department, when the reagent lot or reagent lot changes Comparability test or parallel test is performed to determine whether the results between lots are reliable. The most commonly used standard domestic laboratories is to obtain %difference from the difference in results between two lots of reagents, and then many laboratories are set the standard to less than 20% at low concentrations and less than 10% at medium and high concentrations. If the range is deviated from the standard, the test is considered failed and it is repeated until the result falls within the standard range. In this study, several tests are selected that are performed in nuclear medicine in-vitro laboratories to analyze parallel test results and to establish criteria for customized percent difference for each test. Materials and Methods From January to November 2018, the result of parallel test for reagent lot change is analyzed for 7 items including thyroid-stimulating hormone (TSH), free thyroxine (FT4), carcinoembryonic antigen (CEA), CA-125, prostate-specific antigen (PSA), HBs-Ab and Insulin. The RIA-MAT 280 system which adopted the principle of IRMA is used for TSH, FT4, CEA, CA-125 and PSA. TECAN automated dispensing equipment and GAMMA-10 is used to measure insulin test. For the test of HBs-Ab, HAMILTON automated dispensing equipment and Cobra Gamma ray measuring instrument are used. Separate reagent, customized calibrator and quality control materials are used in this experiment. Results 1. TSH [%diffrence Max / Mean / Median] (P-value by t-test > 0.05) C-1(low concentration) [14.8 / 4.4 / 3.7 / 0.0 ] C-2(middle concentration) [10.1 / 4.2 / 3.7 / 0.0] 2. FT4 [%diffrence Max / Mean / Median] (P-value by t-test > 0.05) C-1(low concentration) [10.0 / 4.2 / 3.9 / 0.0] C-2(high concentration) [9.6 / 3.3 / 3.1 / 0.0 ] 3. CA-125 [%diffrence Max / Mean / Median] (P-value by t-test > 0.05) C-1(middle concentration) [9.6 / 4.3 / 4.3 / 0.3] C-2(high concentration) [6.5 / 3.5 / 4.3 / 0.4] 4. CEA [%diffrence Max / Mean / median] (P-value by t-test > 0.05) C-1(low concentration) [9.8 / 4.2 / 3.0 / 0.0] C-2(middle concentration) [8.7 / 3.7 / 2.3 / 0.3] 5. PSA [%diffrence Max / Mean / Median] (P-value by t-test > 0.05) C-1(low concentration) [15.4 / 7.6 / 8.2 / 0.0] C-2(middle concentration) [8.8 / 4.5 / 4.8 / 0.9] 6. HBs-Ab [%diffrence Max / Mean / Median] (P-value by t-test > 0.05) C-1(middle concentration) [9.6 / 3.7 / 2.7 / 0.2] C-2(high concentration) [8.9 / 4.1 / 3.6 / 0.3] 7. Insulin [%diffrence Max / Mean / Median] (P-value by t-test > 0.05) C-1(middle concentration) [8.7 / 3.1 / 2.4 / 0.9] C-2(high concentration) [8.3 / 3.2 / 1.5 / 0.1] In some low concentration measurements, the percent difference is found above 10 to nearly 15 percent in result of target value calculated at a lower concentration. In addition, when the value is measured after Standard level 6, which is the highest value of reagents in the dispensing sequence, the result would have been affected by a hook effect. Overall, there was no significant difference in lot change of quality control material (p-value>0.05). Conclusion Variations between reagent lots are not large in immunoradiometric assays. It is likely that this is due to the selection of items that have relatively high detection rate in the immunoradiometric method and several remeasurements. In most test results, the difference was less than 10 percent, which was within the standard range. TSH control level 1 and PSA control level 1, which have low concentration target value, exceeded 10 percent more than twice, but it did not result in a value that was near 20 percent. As a result, it is required to perform a longer period of observation for more homogenized average results and to obtain laboratory-specific acceptance criteria for each item. Also, it is advised to study observations considering various variables.