An integrated stack capacitor (ISC) solution, which can effectively suppress power noise in high frequency bands, is introduced. The basic structure of the ISC is a vertical cylinder array consisting of many capacitive vias. The proposed ISC shows high capacitance density compared to the existing silicon capacitors. In this study, the power integrity (PI) performance gain of the proposed ISC solution was analyzed by applying it to the advanced package platforms such as 2.5D silicon interposer, fanout (FO) package, RDL interposer, and substrate based chiplet. Based on 3D wafer on wafer (WoW) technology, ISC is a not only 2.5D silicon interposer for high performance computing (HPC) and server that operates with high power, but also a novel silicon capacitor solution that can be applied to substrate and fanout packages for mobile and automotive.

An ensemble classifier is a method that combines output of multiple classifiers. It has been widely accepted that ensemble classifiers can improve the prediction accuracy. Recently, ensemble techniques have been successfully applied to the bankruptcy prediction. Bagging and random subspace are the most popular ensemble techniques. Bagging and random subspace have proved to be very effective in improving the generalization ability respectively. However, there are few studies which have focused on the integration of bagging and random subspace. In this study, we proposed a new hybrid ensemble model to integrate bagging and random subspace method using genetic algorithm for improving the performance of the model. The proposed model is applied to the bankruptcy prediction for Korean companies and compared with other models in this study. The experimental results showed that the proposed model performs better than the other models such as the single classifier, the original ensemble model and the simple hybrid model.

Ensemble classification is an approach that combines individually trained classifiers in order to improve prediction accuracy over individual classifiers. Ensemble techniques have been shown to be very effective in improving the generalization ability of the classifier. But base classifiers need to be as accurate and diverse as possible in order to enhance the generalization abilities of an ensemble model. Bagging is one of the most popular ensemble methods. In bagging, the different training data subsets are randomly drawn with replacement from the original training dataset. Base classifiers are trained on the different bootstrap samples. In this study we proposed a new bagging variant ensemble model, Randomized Bagging (RBagging) for improving the standard bagging ensemble model. The proposed model was applied to the bankruptcy prediction problem using a real data set and the results were compared with those of the other models. The experimental results showed that the proposed model outperformed the standard bagging model.

Ensemble classification combines individually trained classifiers to obtain more accurate predictions than individual classifiers alone. Ensemble techniques are very useful for improving the generalizability of the classifier. Bagging is the method used most commonly for constructing ensemble classifiers. In bagging, different training data subsets are drawn randomly with replacement from the original training dataset. Base classifiers are trained on these different bootstrap samples. Instance selection is used to select critical instances while deleting and removing irrelevant and harmful instances from the original set. Although instance selection and bagging have proven effective in many data mining applications, few studies have considered the integration of the two. This study proposes a new method to integrate instance selection and bagging ensemble using genetic algorithms to improve the performance of the model. A genetic algorithm is used to select optimal or near-optimal instances to be used as input data by the bagging model. This study applies the proposed model to a bankruptcy-prediction problem. The experimental results show that the proposed model outperformed the other models.

Ensemble classification involves combining multiple classifiers to obtain more accurate predictions than those obtained using individual models. Ensemble techniques are known to be very useful in improving the generalization ability of a classifier. The random subspace ensemble technique is a simple but effective method of constructing ensemble classifiers, in which some features are randomly drawn from all features of each classifier in the ensemble. Recently, ensemble techniques have been successfully applied to bankruptcy prediction, but few studies have incorporated heterogeneous random subspace models for bankruptcy prediction. A heterogeneous ensemble is a set of base classifiers created using different algorithms, while a homogeneous ensemble uses only one algorithm to create base classifiers. In this study, we applied a heterogeneous random subspace model to the bankruptcy prediction problem. We also developed a method for optimizing the heterogeneous random subspace ensemble model, using a genetic algorithm to optimize its classifier subsets. We applied the proposed model to a bankruptcy prediction problem using a real dataset from Korean companies. The experimental results confirmed that the proposed model outperformed other models.

앙상블 학습 기법은 개별 모형보다 더 좋은 예측 성과를 얻기 위해 다수의 분류기를 결합하는 것으로 예측 성과를 향상시키는데에 매우 유용한 것으로 알려져 있다. 배깅은 단일 분류기의 예측 성과를 향상시키는 대표적인 앙상블 기법중의 하나이다. 배깅은 원 학습 데이터로부터 부트스트랩 샘플링 방법을 통해 서로 다른 학습 데이터를 추출하고, 각각의 부트스트랩 샘플에 대해 학습 알고리즘을 적용하여 서로 다른 다수의 기저 분류기들을 생성시키게 되며, 최종적으로 서로 다른 분류기로부터 나온 결과를 결합하게 된다. 배깅에서 부트스트랩 샘플은 원 학습 데이터로부터 램덤하게 추출한 샘플로 각각의 부트스트랩 샘플이 동일한 정보를 가지고 있지는 않으며 이로 인해 배깅 모형의 성과는 편차가 발생하게 된다. 본 논문에서는 이와 같은 부트스트랩 샘플을 최적화함으로써 표준 배깅 앙상블의 성과를 개선시키는 새로운 방법을 제안하였다. 제안한 모형에서는 앙상블 모형의 성과를 개선시키기 위해 부트스트랩 샘플링을 최적화하였으며 이를 위해 유전자 알고리즘이 활용되었다. 본 논문에서는 제안한 모형을 국내 부도 예측 문제에 적용해 보았으며, 실험 결과 제안한 모형이 우수한 성과를 보였다. Ensemble classification involves combining multiple classifiers to obtain more accurate predictions than those obtained using individual models. Ensemble learning techniques are known to be very useful for improving prediction accuracy. Bagging is one of the most popular ensemble learning techniques. Bagging has been known to be successful in increasing the accuracy of prediction of the individual classifiers. Bagging draws bootstrap samples from the training sample, applies the classifier to each bootstrap sample, and then combines the predictions of these classifiers to get the final classification result. Bootstrap samples are simple random samples selected from the original training data, so not all bootstrap samples are equally informative, due to the randomness. In this study, we proposed a new method for improving the performance of the standard bagging ensemble by optimizing bootstrap samples. A genetic algorithm is used to optimize bootstrap samples of the ensemble for improving prediction accuracy of the ensemble model. The proposed model is applied to a bankruptcy prediction problem using a real dataset from Korean companies. The experimental results showed the effectiveness of the proposed model.

Ensemble classification is to utilize multiple classifiers instead of using a single classifier. Recently ensemble classifiers have attracted much attention in data mining community. Ensemble learning techniques has been proved to be very useful for improving the prediction accuracy. Bagging, boosting and random subspace are the most popular ensemble methods. In random subspace, each base classifier is trained on a randomly chosen feature subspace of the original feature space. The outputs of different base classifiers are aggregated together usually by a simple majority vote. In this study, we applied the random subspace method to the bankruptcy problem. Moreover, we proposed a method for optimizing the random subspace ensemble. The genetic algorithm was used to optimize classifier subset of random subspace ensemble for bankruptcy prediction. This paper applied the proposed genetic algorithm based random subspace ensemble model to the bankruptcy prediction problem using a real data set and compared it with other models. Experimental results showed the proposed model outperformed the other models.

기업의 부도 예측은 재무 및 회계 분야에서 매우 중요한 연구 주제이다. 기업의 부도로 인해 발생하는 비용이 매우 크기 때문에 부도 예측의 정확성은 금융기관으로서는 매우 중요한 일이다. 최근에는 여러 개의 모형을 결합하는 앙상블 모형을 부도 예측에 적용해 보려는 연구가 큰 관심을 끌고 있다. 앙상블 모형은 개별 모형보다 더 좋은 성과를 내기 위해 여러 개의 분류기를 결합하는 것이다. 이와 같은 앙상블 분류기는 분류기의 일반화 성능을 개선하는 데 매우 유용한 것으로 알려져 있다. 본 논문은 부도 예측 모형의 성과 개선에 관한 연구이다. 이를 위해 사례 선택(Instance Selection)을 활용한 배깅(Bagging) 모형을 제안하였다. 사례 선택은 원 데이터에서 가장 대표성 있고 관련성 높은 데이터를 선택하고 예측 모형에 악영향을 줄 수 있는 불필요한 데이터를 제거하는 것으로 이를 통해 예측 성과 개선도 기대할 수 있다. 배깅은 학습데이터에 변화를 줌으로써 기저 분류기들을 다양화시키는 앙상블 기법으로 단순하면서도 성과가 매우 좋은 것으로 알려져 있다. 사례 선택과 배깅은 각각 모형의 성과를 개선시킬 수 있는 잠재력이 있지만 이들 두 기법의 결합에 관한 연구는 아직까지 없는 것이 현실이다. 본 연구에서는 부도 예측 모형의 성과를 개선하기 위해 사례 선택과 배깅을 연결하는 새로운 모형을 제안하였다. 최적의 사례 선택을 위해 유전자 알고리즘이 사용되었으며, 이를 통해 최적의 사례 선택 조합을 찾고 이 결과를 배깅 앙상블 모형에 전달하여 새로운 형태의 배깅 앙상블 모형을 구성하게 된다. 본 연구에서 제안한 새로운 앙상블 모형의 성과를 검증하기 위해 ROC 커브, AUC, 예측정확도 등과 같은 성과지표를 사용해 다양한 모형과 비교 분석해 보았다. 실제 기업데이터를 사용해 실험한 결과 본 논문에서 제안한 새로운 형태의 모형이 가장 좋은 성과를 보임을 알 수 있었다. Predicting corporate failure has been an important topic in accounting and finance. The costs associated with bankruptcy are high, so the accuracy of bankruptcy prediction is greatly important for financial institutions. Lots of researchers have dealt with the topic associated with bankruptcy prediction in the past three decades. The current research attempts to use ensemble models for improving the performance of bankruptcy prediction. Ensemble classification is to combine individually trained classifiers in order to gain more accurate prediction than individual models. Ensemble techniques are shown to be very useful for improving the generalization ability of the classifier. Bagging is the most commonly used methods for constructing ensemble classifiers. In bagging, the different training data subsets are randomly drawn with replacement from the original training dataset. Base classifiers are trained on the different bootstrap samples. Instance selection is to select critical instances while deleting and removing irrelevant and harmful instances from the original set. Instance selection and bagging are quite well known in data mining. However, few studies have dealt with the integration of instance selection and bagging. This study proposes an improved bagging ensemble based on instance selection using genetic algorithms (GA) for improving the performance of SVM. GA is an efficient optimization procedure based on the theory of natural selection and evolution. GA uses the idea of survival of the fittest by progressively accepting better solutions to the problems. GA searches by maintaining a population of solutions from which better solutions are created rather than making incremental changes to a single solution to the problem. The initial solution population is generated randomly and evolves into the next generation by genetic operators such as selection, crossover and mutation. The solutions coded by strings are evaluated by the fitness function. The proposed model consists of two phases: GA based Instance Selection and Instance based Bagging. In the first phase, GA is used to select optimal instance subset that is used as input data of bagging model. In this study, the chromosome is encoded as a form of binary string for the instance subset. In this phase, the population size was set to 100 while maximum number of generations was set to 150. We set the crossover rate and mutation rate to 0.7 and 0.1 respectively. We used the prediction accuracy of model as the fitness function of GA. SVM model is trained on training data set using the selected instance subset. The prediction accuracy of SVM model over test data set is used as fitness value in order to avoid overfitting. In the second phase, we used the optimal instance subset selected in the first phase as input data of bagging model. We used SVM model as base classifier for bagging ensemble. The majority voting scheme was used as a combining method in this study. This study applies the proposed model to the bankruptcy prediction problem using a real data set from Korean companies. The research data used in this study contains 1832 externally non-audited firms which filed for bankruptcy (916 cases) and non-bankruptcy (916 cases). Financial ratios categorized as stability, profitability, growth, activity and cash flow were investigated through literature review and basic statistical methods and we selected 8 financial ratios as the final input variables. We separated the whole data into three subsets as training, test and validation data set. In this study, we compared the proposed model with several comparative models including the simple individual SVM model, the simple bagging model and the instance selection based SVM model. The McNemar tests were used to examine whether the proposed model significantly outperforms the other models. The experimental results show that the proposed model outperforms the other models.

This paper proposes an effective approach to suppress vertical electromagnetic (EM) coupling in multilayer packages operating at GHz frequencies. In the case of packages with embedded actives where there are large apertures (die sized) in the metal planes and cavities in dielectric layers to accommodate the chips, the effect of EM field coupling is significant. The method involves EM band-gap structures for suppressing vertical coupling and the isolation band can be tuned over different frequency ranges. In addition, this paper puts forth a methodology for predicting the frequency range of the isolation band achieved by the coupling suppression technique. The proposed methodology is demonstrated through simulations and measurements.

분류기의 앙상블 학습은 여러 개의 서로 다른 분류기들의 조합을 통해 만들어진다. 앙상블 학습은 기계학습 분야에서 많은 관심을 끌고 있는 중요한 연구주제이며 대부분의 경우에 있어서 앙상블 모형은 개별 기저 분류기보다 더 좋은 성과를 내는 것으로 알려져 있다. 본 연구는 부도 예측 모형의 성능개선에 관한 연구이다. 이를 위해 본 연구에서는 단일 모형으로 그 우수성을 인정받고 있는 SVM을 기저 분류기로 사용하는 앙상블 모형에 대해 고찰하였다. SVM 모형의 성능 개선을 위해 bagging과 random subspace 모형을 부도 예측 문제에 적용해 보았으며 bagging 모형과 random subspace 모형의 성과 개선을 위해 bagging과 random subspace의 통합 모형을 제안하였다. 제안한 모형의 성과를 검증하기 위해 실제 기업의 부도 예측 데이터를 사용하여 실험하였고, 실험 결과 본 연구에서 제안한 새로운 형태의 통합 모형이 가장 좋은 성과를 보임을 알 수 있었다. An ensemble of classifiers is to employ a set of individually trained classifiers and combine their predictions. It has been found that in most cases the ensembles produce more accurate predictions than the base classifiers. Combining outputs from multiple classifiers, known as ensemble learning, is one of the standard and most important techniques for improving classification accuracy in machine learning. An ensemble of classifiers is efficient only if the individual classifiers make decisions as diverse as possible. Bagging is the most popular method of ensemble learning to generate a diverse set of classifiers. Diversity in bagging is obtained by using different training sets. The different training data subsets are randomly drawn with replacement from the entire training dataset. The random subspace method is an ensemble construction technique using different attribute subsets. In the random subspace, the training dataset is also modified as in bagging. However, this modification is performed in the feature space. Bagging and random subspace are quite well known and popular ensemble algorithms. However, few studies have dealt with the integration of bagging and random subspace using SVM Classifiers, though there is a great potential for useful applications in this area. The focus of this paper is to propose methods for improving SVM performance using hybrid ensemble strategy for bankruptcy prediction. This paper applies the proposed ensemble model to the bankruptcy prediction problem using a real data set from Korean companies.

This paper presents thin-film high-rejection bandpass filters integrated into organic substrate technology called RXP. A circuit topology based on unique resonator and grounding inductor is proposed. The proposed circuit provides design freedom of locating transmission zeros at lower and upper stopbands relative to the center frequency. About 2.4 GHz narrow band (<200 MHz) and 5 GHz wide band (>1 GHz) bandpass filters have been implemented in four-metal layer with 0.191 mm thin RXP substrate. The measured results show good agreement with simulations.

This paper presents, for the first time, characterization results of next generation dielectric core and build up material called RXP, which has low dielectric constant (2.93-3.48) and low loss tangent (0.0037-0.006) up to 110 GHz. Unlike LCP, this material can be made ultra-thin with low processing temperature and is ideally suited for mobile applications. Causal models suitable for high frequency applications have been extracted by measuring the response of cavity resonators using vector network analyzer and surface profiler.

Interposer technology has evolved from ceramic to organic materials and most recently to silicon. Organic substrates exhibit poor dimensional stability, thus requiring large capture pads which make them unsuitable for very high I/Os with fine pitch interconnections. Therefore, there has been a trend to develop silicon interposers. Silicon interposers however, suffer in two ways; 1) they are expensive to process due to the need for electrical insulation around via walls, and 2) they are limited in size by the silicon wafer from which they originate. In this paper, glass is proposed as a superior alternative interposer technology to address the limitations of both silicon and organic interposers. The inherent electrical properties of glass, together with large area panel size availability, make it superior compared to organic and silicon-based interposers. Glass however, is not without its challenges. It suffers in two ways: 1) formation of vias at low cost, and 2) its lower thermal conductivity compared to silicon. This research explores glass as an interposer material, and addresses the above key challenges in through package via (TPV) formation and subsequent low cost and large area metallization to achieve very high I/Os at fine pitch.

A lumped-element bandpass filter based on new RXP ultra-thin organic technology with enhanced stopband rejections is proposed in this paper. The design is based on a third-order capacitively-coupled resonator circuit with unique resonator and ground inductor. For demonstration of the proposed circuit and RXP technology, a 5 GHz bandpass filter has been implemented in a four-metal layer 0.191 mm thin RXP substrate. The measured results have a good agreement with the simulation and show that insertion loss is less than 1.24 dB with larger than 1 GHz bandwidth and sharp rejections at both low and high stopband.

This paper presents the design of filters and antennas in advanced polymers based on a new material called RXP. Integration capability of the RXP and the performance of WLAN filter design have been verified through the measurement data presented in this paper. Simulated results of 60 GHz filters and antennas are also included in this paper. RXP provides low cost and promising high performance advanced polymer solution for wireless applications operating around microwave and millimeter frequencies.

This paper presents the integration of WLAN (2.4 and 5GHz) bandpass filters in glass interposer using through-package vias. The filters include novel embedded passive components such as stitched capacitors with reduced shunt parasitics and via-based inductors that provide area reduction. The filters designed for 2.4 GHz showed an insertion loss of less than 2dB and better than 15dB return loss, while the 5GHz filters showed an insertion loss of less than 1dB with better than 20dB return loss. Stop-band rejection of over 35dB was observed at 2.2 GHz on the 2.4 GHz bandpass filters. The measured results showed good agreement with the simulated values and indicated that the performance on glass interposer closely matches the performance of the more expensive high resistivity silicon with similar properties.

This paper presents for the first time a novel manufacturing-compatible organic substrate and interconnect technology using ultra-thin chip-last embedded active and passive components for digital, analog, MEMS, RF, microwave and millimeter wave applications. The architecture of the platform consists of a low-CTE thin core and minimum number of thin build up organic dielectric and conductive layers. This organic substrate is based on a new generation of low-loss and thermally-stable thermosetting polymers (RXP-1 and RXP-4). Unlike LCP- and Teflon-based materials, the RXP material system is fully compatible with conventional FR-4 manufacturing processes. Ultra-thin silicon test die (55µm thick) has been embedded in a 60µm deep cavity with a 6-metal layer RXP substrate and a total thickness of 0.22mm. The embedded IC is interconnected to the substrate by ultra-fine pitch Cu-to-Cu bonding with polymer adhesives. This novel interconnection process performed at 180°C, has passed 1,000 thermal shock cycles in reliability testing. Because of manufacturing process simplicity and unparalleled set of benefits, the chip-last technology described in this paper provides the benefits of chip-first without its disadvantages and thus enables highly miniaturized, multi-band, high performance 3D modules by stacking embedded 3D ICs or packages with embedded actives, passives and MEMS devices.