International Telemetering Conference Proceedings / October 28-31, 1996 / Town and Country Hotel and Convention Center, San Diego, California / This paper will discuss the design and use of Optimal Ratio Combiners in modern telemetry applications. This will include basic design theory, operational setups, and various types of combiner configurations. The paper will discuss the advantages of pre-detection vs. post-detection combining. Finally, the paper will discuss modern design techniques.

This thesis aims to design transformer-based balanced passive components with high performance and compact size using CMOS and printed-circuit¡Vboard (PCB) technologies. A CMOS parallel-combining transformer (PCT) incorporating a planar trifilar transformer is presented to realize power combining and impedance transformation at the same time. In addition, a CMOS Wilkinson power combiner with a planar bifilar transformer is proposed to enhance isolation between two combining ports. Several transformer coupled balun designs with an overlay winding structure are carried out on FR4 and Duroid substrates. These designs uses a rather symmetric layout to achieve a superior balance performance and a multilayer configuration to create the transmission zeros in the out-of-band response. Finally, a CMOS transformer balun is implemented with a bandpass filter passband which is designed according to the coupled resonator filter theory. This passband can restrict the bandwidth usage for the balun to improve the common-mode rejection ratio (CMRR) within the passband.

Purpose of this study is to use adjustable platform for development of fiber diameter and fiber fusion from 125£gm to 50£gm,125£gm to 40£gm different taper single-mode fiber, the use of different process parameters in the design of fire taper models and parameters with the molten zone, the minimum loss of 0.77dB.After the large diameter quartz capillary tube inserted into some optical fiber taper changes made after the n¡Ñ1 optical combiner, and adjusted through the design taper model parameters reduce the firepower some fiber transmission loss. In recent years, the application of high-power fiber laser in various industries is widely spreading and exploring. Within the system, one of critical component is the fiber pump combiner. We are currently investigating the fabrication of 7x1 Combiner with both single mode and multimode fibers and the characterization of combiner efficiency as a function of processing. The development of 7¡Ñ1 single-mode fiber combiner loss of about 1.5dB and 7 ¡Ñ 1 multimode fiber combiner loss of about 1.4dB

This thesis at first introduces planar transformers and their applications to RF circuits. A method is given to validate the impedance transformation and calculate the insertion loss for a transformer with unequal turn windings. A physical lumped-element model has been established for a planar transformer-based balun implemented using integrated passive device technology. This thesis next presents CMOS power combiner designs with two differential ports and one single-ended port. Such designs equivalently integrate two baluns and one single-ended power combiner into a single circuit and therefore have the advantages of saving chip area and reducing insertion loss. In addition, the designs consider an annular ground structure to achieve lower insertion loss or better balance property. Finally, this thesis presents a balun using a stacked coil transformer in an organic build-up substrate. For this purpose, a complete design flow and an impedance matching technique have been given.

International Telemetering Conference Proceedings / October 17-20, 1994 / Town & Country Hotel and Conference Center, San Diego, California / With the increased concerns for reducing cost and improving reliability in today's telemetry systems, many users are employing simulation and automation to guarantee reliable telemetry systems operation. This places an increased demand on the remote capabilities of the equipment used in the telemetry system. Furthermore, emphasis has been placed on the ability to decrease the space and power consumption of the telemetry system to facilitate transportability of the a single telemetry system to multiple sites. Finally, today's telemetry systems demand that all equipment provide multiple functions to provide the maximum performance for the lowest system cost.

The work done in this thesis focuses on the design of a square-shaped 20-way metamaterial power divider which is fabricated in microstrip technology and operates at 1 GHz. The divider comprises 12 square-shaped left-handed unit cells and 13 square-shaped right-handed unit cells, and these unit cells have the same size and are placed in a checker-board tessellation, where the left-handed unit cells are connected only to right-handed unit cells and vice versa. The divider is based upon the infinite wavelength phenomenon in two-dimensions, and this means that the insertion phase between any two ports of the left-handed unit cell is equal, but with opposite sign, to that of the right-handed unit cell. The divider gives an equal-amplitude equal-phase power division from the central input port to the output ports which are located on a straight line on each side. Thus, it is convenient to integrate with, or interconnect to, other planar circuits in a system, such as power amplifier modules. The design concept can be extended to an N-way power divider, where N = 4n and n is an odd integer.

Smart antenna technology is a promising means to overcome signal impairments in wireless personal communications. When spatial signal processing achieved through smart antennas is combined with temporal signal processing, the space-time processing can mitigate interference and multipath to yield higher network capacity, coverage, and quality. In this dissertation, we propose a dual smart antenna system incorporated into handsets for the third generation wireless personal communication systems in which the two antennas are separated by a quarter wavelength (3.5 cm). We examine the effectiveness of a dual smart antenna system with diversity and adaptive combining schemes and propose a new combining scheme called hybrid combining. The proposed hybrid combiner combines diversity combiner and adaptive combiner outputs using maximal ratio combining (MRC). Since these diversity combining and adaptive combining schemes exhibit somewhat opposite and complementary characteristics, the proposed hybrid combining scheme aims to exploit the advantages of the two schemes. To model dual antenna signals, we consider three channel models: loosely correlated fading channel model (LCFCM), spatially correlated fading channel model (SCFCM), and envelope correlated fading channel model (ECFCM). Each antenna signal is assumed to have independent Rayleigh fading in the LCFCM. In the SCFCM, each antenna signal is subject to the same Rayleigh fading, but is different in the phase due to a non-zero angle of arrival (AOA). The LCFCM and the SCFCM are useful to evaluate the upper and the lower bounds of the system performance. To model the actual channel of dual antenna signals lying in between these two channel models, the ECFCM is considered. In this model, two Rayleigh fading antenna signals for each multipath are assumed to have an envelope correlation and a phase difference due to a non-zero AOA. To obtain the channel profile, we adopted not only the geometrically based single bounce (GBSB) circular and elliptical models, but also the International Telecommunication Union (ITU) channel model. In this dissertation, we also propose a new generalized selection combining (GSC) method called minimum selection GSC (MS-GSC) and an adaptive rake combining scheme to reduce the power consumption of mobile rake receivers. The proposed MS-GSC selects a minimum number of branches as long as the combined SNR is maintained larger than a given threshold. The proposed adaptive rake combining scheme which dynamically determines the threshold values is applicable to the three GSC methods: the absolute threshold GSC, the normalized threshold GSC, and the proposed MS-GSC. Through simulation, we estimated the effectiveness of the proposed scheme for a mobile rake receiver for a wideband CDMA system. We also suggest a new power control strategy to maximize the benefit of the proposed adaptive scheme. / Ph. D.

Abstract Existing classification analysis techniques (e.g., decision tree induction, backpropagation neural network, k-nearest neighbor classification, etc.) generally exhibit satisfactory classification effectiveness when dealing with data with non-skewed class distribution. However, real-world applications (e.g., churn prediction and fraud detection) often involve highly skewed data in decision outcomes (e.g., 2% churners and 98% non-churners). Such a highly skewed class distribution problem, if not properly addressed, would imperil the resulting learning effectiveness and might result in a ¡§null¡¨ prediction system that simply predicts all instances as having the majority decision class as the training instances (e.g., predicting all customers as non-churners). In this study, we extended the multi-classifier class-combiner approach and proposed a clustering-based multi-classifier class-combiner technique to address the highly skewed class distribution problem in classification analysis. In addition, we proposed four distance-based methods for selecting a subset of instances having the majority decision class for lowering the degree of skewness in a data set. Using two real-world datasets (including mortality prediction for burn patients and customer loyalty prediction), empirical results suggested that the proposed clustering-based multi-classifier class-combiner technique generally outperformed the traditional multi-classifier class-combiner approach and the four distance-based methods. Keywords: Data Mining, Classification Analysis, Skewed Class Distribution Problem, Decision Tree Induction, Multi-classifier Class-combiner Approach, Clustering-based Multi-classifier Class-combiner Approach

The goal of this dissertation is to improve financial statement fraud detection using a cross-functional research approach. The efficacy of financial statement fraud detection depends on the classification algorithms and the fraud predictors used and how they are combined. Essay I introduces IMF, a novel combiner method classification algorithm. The results show that IMF performs well relative to existing combiner methods over a wide range of domains. This research contributes to combiner method research and, thereby, to the broader research stream of ensemble-based classification and to classification algorithm research in general. Essay II develops three novel fraud predictors: total discretionary accruals, meeting or beating analyst forecasts and unexpected employee productivity. The results show that the three variables are significant predictors of fraud. Hence Essay II provides insights into (1) conditions under which fraud is more likely to occur (total discretionary accruals is high), (2) incentives for fraud (firms desire to meet or beat analyst forecasts), and (3) how fraud is committed and can be detected (revenue fraud detection using unexpected employee productivity). This essay contributes to confirmatory fraud predictor research, which is a sub-stream of research that focuses on developing and testing financial statement fraud predictors. Essay III compares the utility of artifacts developed in the broader research streams to which the first two essays contribute, i.e., classification algorithm and fraud predictor research in detecting financial statement fraud. The results show that logistic regression and SVM perform well, and that out of 41 variables found to be good predictors in prior fraud research, only six variables are selected by three or more classifiers: auditor turnover, Big 4 auditor, accounts receivable and the three variables introduced in Essay II. Together, the results from Essay I and Essay III show that IMF performs better than existing combiner methods in a wide range of domains and better than stacking, an ensemble-based classification algorithm, in fraud detection. The results from Essay II and Essay III show that the three predictors created in Essay II are significant predictors of fraud and, when evaluated together with 38 other predictors, provide utility to classification algorithms.

Classifier ensembling research has been one of the most active areas of machine learning for a long period of time. The main aim of generating combined classifier ensembles is to improve the prediction accuracy in comparison to using an individual classifier. A combined classifiers ensemble can improve the prediction results by compensating for the individual classifier weaknesses in certain areas and benefiting from better accuracy of the other ensembles in the same area. In this thesis, different algorithms are proposed for designing classifier ensemble combiners. The existing methods such as averaging, voting, weighted average, and optimised weighted method does not increase the accuracy of the combiner in comparison to the proposed advanced methods such as genetic programming and the coalition method. The different methods are studied in detail and analysed using different databases. The aim is to increase the accuracy of the combiner in comparison to the standard stand-alone classifiers. The proposed methods are based on generating a combiner formula using genetic programming, while the coalition is based on estimating the diversity of the classifiers such that a coalition is generated with better prediction accuracy. Standard accuracy measures are used, namely accuracy, sensitivity, specificity and area under the curve, in addition to training error accuracies such as the mean square error. The combiner methods are compared empirically with several stand-alone classifiers using neural network algorithms. Different types of neural network topologies are used to generate different models. Experimental results show that the combiner algorithms are superior in creating the most diverse and accurate classifier ensembles. Ensembles of the same models are generated to boost the accuracy of a single classifier type. An ensemble of 10 models of different initial weights is used to improve the accuracy. Experiments show a significant improvement over a single model classifier. Finally, two combining methods are studied, namely the genetic programming and coalition combination methods. The genetic programming algorithm is used to generate a formula for the classifiers’ combinations, while the coalition method is based on a simple algorithm that assigns linear combination weights based on the consensus theory. Experimental results of the same databases demonstrate the effectiveness of the proposed methods compared to conventional combining methods. The results show that the coalition method is better than genetic programming.