• Journal of the Economic Geographical Society of Korea
• /
• v.13 no.4
• /
• pp.641-662
• /
• 2010

Through the comparative studies on Korean and Chinese ethnobanks in the U.S., this research finds the difference in the social capital between Korean and Chinese American economy and reinterprets the social capital in overseas Chinese. In many ethnic studies, the ethnobanks in the U.S. are thought as the main drive of the economic prosperity in the ethnic communities. For the success of these ethnobanks, the relationship banking of these ethnobanks are contrasted with the arms-length banking of the mainstream U.S. banks. However there are the differences both in performance and business strategies even among different ethnobanks. Contrary to the Korean ethnic banks having much portion of business loans, Chinese ethnic banks have stressed real-estate loans in their loan portfolio. This research regards the differences of business strategies and performance of these ethnobanks as the result of the different social, cultural contexts in each ethnic community. From these finings, we can improve the understanding of the concept of social capital in overseas Chinese economy.

• Journal of the Korean Data Analysis Society
• /
• v.20 no.6
• /
• pp.3037-3048
• /
• 2018

This study tried to test the result of recent study reporting that there is no the effect of cross-subsidization by applying it to the Korean banks. Besides, the study also did the research to test the previous study that diversification toward non-interest operating sector is related with the motivation of reduction of bank risk. Main results are as follows. First, there is the effect of cross-subsidization between non-interest operating sector and interest one. The higher non-interest profit to total asset is, the lower loan growth and provision to total loan are. Second, the effect of cross-subsidization is related with the fee revenue to operating revenue rather than the fiduciary revenue to operating revenue because high fee revenue results in shrinking the lending business. Third, diversification toward non-interest sector is associated with reducing risk, and risk is not increased when the high performance of non-interest sector is occurred. This evidence is not line with DeYoung, Torna (2013) arguing that non-interest operating activities lead to increasing risk. Finally, Korean banks have to expand the activities in non-interesting sector focusing on boosting fee revenue to increase the effect of cross-subsidization against the possibility of cutting net interest margin.

• Journal of Internet Computing and Services
• /
• v.25 no.1
• /
• pp.57-67
• /
• 2024

This research investigates training methods for large language models to accurately identify sentiments and comprehend information about increasing and decreasing fluctuations in the financial domain. The main goal is to identify suitable datasets that enable these models to effectively understand expressions related to financial increases and decreases. For this purpose, we selected sentences from Wall Street Journal that included relevant financial terms and sentences generated by GPT-3.5-turbo-1106 for post-training. We assessed the impact of these datasets on language model performance using Financial PhraseBank, a benchmark dataset for financial sentiment analysis. Our findings demonstrate that post-training FinBERT, a model specialized in finance, outperformed the similarly post-trained BERT, a general domain model. Moreover, post-training with actual financial news proved to be more effective than using generated sentences, though in scenarios requiring higher generalization, models trained on generated sentences performed better. This suggests that aligning the model's domain with the domain of the area intended for improvement and choosing the right dataset are crucial for enhancing a language model's understanding and sentiment prediction accuracy. These results offer a methodology for optimizing language model performance in financial sentiment analysis tasks and suggest future research directions for more nuanced language understanding and sentiment analysis in finance. This research provides valuable insights not only for the financial sector but also for language model training across various domains.

• Journal of Digital Convergence
• /
• v.13 no.8
• /
• pp.123-131
• /
• 2015

Modern is the era of convergence. In the complex position of the company to survive in the rapid environmental changes it can not be a mandatory fusion rather than choice. Wind convergence also in the art can be said to castration. This phenomenon can be found in many real life. Most real-life contact with the service sector is no exception. Due to a number of banks and insurance combined Banca named fusion of French banking and insurance business are integrated. Even recently, the situation has been provided by Investment consulting, equity trading services within bank branches. In this study, based on the need for convergence of corporate businesses convergence factors examined through regression analysis on whether any impact on the financial aptitude. External environmental factors that make up the fusion research factors internal environmental factors, both core competencies were factors to affect financial performance.

• Journal of Advanced Navigation Technology
• /
• v.6 no.2
• /
• pp.138-150
• /
• 2002

This paper addresses the problem of combining adaptive polarization processing and space-time processing for further performance improvement of radar target detection in clutter and Jammer environments. Since the most straightforward cascade combinations have quite limited performance improvement potentials, we focus on the development of adaptive processing in the joint polarization-space-time domain. Unlike a direct extension of some existing space-time processing algorithms to the joint domain, the processing algorithm developed in this paper does not need a potentially costly polarization filter bank to cover the unknown target polarization parameter. The performance of the new algorithm is derived and evaluated in terms of the probability of detection and the probability of false alarm, and it is compared with other algorithms that do not utilize the polarization information or assume that the target polarization is known.

• Publications of The Korean Astronomical Society
• /
• v.22 no.3
• /
• pp.63-73
• /
• 2007

In this paper, we introduce the performance test results of digital data processing system for KVN (Korean VLBI Network). The digital data processing system for KVN consists of DAS (Data Acquisition System) and high-speed recorder which called Mark5B system. DAS system performs the digitalization of analog radio signal through ADS-1000 gigabit sampler with 1 Gsps/2-bit and process the digital filtering of digital signal. Mark5B system records the output data of DFB (Digital Filter Bank) with about 1 Gbps. In this paper, we carried out the preliminary evaluation experiments of the KVN digital data processing system connected between DAS system and Mark5B with VSI (VLBI Standard Interface) interface which is designed for compatible in each VLBI system. We first performed all of the KVN digital data processing system connected by VSI interface in the world. In factory inspection phase, we found that the DAS system has a memory read/write error in DSM (Digital Spectrometer) by analyzing the recorded data in Mark5B system. We confirmed that the DSM memory error has been correctly solved by comparing DSM results with Mark5B results. The effectiveness of KVN digital data processing system has been verified through the preliminary experiments such as data transmission, recording with VSI interface connection and data analysis between DSM and Mark5B system. In future work, we will perform the real astronomical observation by using the KVN 21m radio telescopes so as to verify its stability and performance.

• Journal of IKEEE
• /
• v.19 no.3
• /
• pp.455-459
• /
• 2015

This paper proposes a dispatch unit of GP-GPU with SIMT architecture to support the acceleration of general-purpose operation as well as graphics processing. If all the information of an operand used instructions issued from the warp scheduler is decoded, an unnecessary operand load occurs, resulting in register loads. To resolve this problem, this paper proposes a method that can reduce the operand load and the load on the resister by decoding only the information of the operand using a pre-decoding method. The operand information from the dispatch unit is passed to the operand selection unit with preventing register bank collisions. Thus the overall performance are improved. In the simulation test, the total clock cycles required by processing 10,000 arbitrary instructions issued from the wrap scheduler using ModelSim SE 10.0b are measured. It shows that the application of the dispatch unit equipped with the pre-decoding function proposed in this paper can make an improvement of about 12% in processing performance compared to the conventional method.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.5
• /
• pp.542-546
• /
• 2008

Recently dual-port SDRAM (DPSDRAM) architecture tailored for dual-processor based mobile embedded systems has been announced where a single memory chip plays the role of the local memories and the shared memory for both processors. In order to maintain memory consistency from simultaneous accesses of both ports, every access to the shared memory should be protected by a synchronization mechanism, which can result in substantial access latency. We propose two optimization techniques by exploiting the communication patterns of target applications: lock-priority scheme and static-copy scheme. Further, by dividing the shared bank into multiple blocks, we allow simultaneous accesses to different blocks thus achieve considerable performance gain. Experiments on a virtual prototyping system show a promising result - we could achieve about 20-50% performance gain compared to the base DPSDRAM architecture.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.6C
• /
• pp.637-644
• /
• 2006

This paper presents performance and implementation comparisons of standard and fast DCT algorithms that are commonly used for subband filter bank in MPEG audio coders. The comparison is made according to the architectural difference of the implementation hardware. Fast DCT algorithms are known to have much less computational complexity than the standard method that involves computing a vector dot product of cosine coefficient. But, due to structural irregularity, fast DCT algorithms require extra cycles to generate the addresses for operands and to realign interim data. When algorithms are implemented using DSP processors that provide special operations such as single-cycle MAC (multiply-accumulate), zero-overhead nested loop, the standard algorithm is more advantageous than the fast algorithms. Also, in case of the finite-precision processing, the error performance of the standard method is far superior to that of the fast algorithms. In this paper, truncation errors and algorithmic suitability are analyzed and implementation results are provided to support the analysis.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.27 no.4
• /
• pp.369-376
• /
• 2016

In this paper, we propose and design a SSD(Simultaneous Single band Duplex) system using 5G(Generation) candidate modulations. Especially, we consider HPA(High Power Amplifier) nonlinearity in the proposed system. And then, we evaluate and analyze performance of the proposed system. As simulation results, performance of SSD-OFDM(Orthogonal Frequency Division Multiplexing), SSD-FMC (Universal Filtered Multi-Carrier), and SSD-FBMC(Filter Bank Multi-Carrier) is severely degraded by HPA nonlinearity. However, performance of SSD-OFDM, SSD-UFMC, and SSD-FBMC is similar in the same condition. That is, OFDM, UFMC, and FBMC have a similar PAPR(Peak to Average Power Ratio) characteristic. Finally, we can confirm that the proposed SSD system can not cancel(SI) self-interference effectively by strong HPA nonlinearity. That is, Reducing PAPR is important in order to avoid effect of HPA nonlinearity in the proposed SSD system.