Note: I request that the MATLAB scripts used in this project not be reproduced on the course website

Size: px
Start display at page:

Download "Note: I request that the MATLAB scripts used in this project not be reproduced on the course website"

Transcription

1 Classification of Breast Tissue Using Electrical Impedance Spectroscopy Michael Nonte Graduate Student Department of Biomedical Engineering University of Wisconsin-Madison Note: I request that the MATLAB scripts used in this project not be reproduced on the course website

2 1. Introduction Breast cancer is typically diagnosed using a tissue biopsy. This is a process by which a sample of suspect tissue is removed from the patient and sent away for histological and chemical analysis. This process typically takes 1 to 2 days, but some samples give inconclusive results and require the patient to have a second biopsy. [1] If a faster method of tissue analysis or a method to determine if a tissue sample is a good candidate for further screening, the cost of screening and stress endured by the patient in awaiting results could be reduced. Electrical Impedance Spectroscopy (EIS) is a method that can used to measure the complex impedance properties of a material. A voltage or current of known frequency and amplitude is applied across the system and the response current or voltage is recorded. The differences in phase and magnitude of the applied stimulus and response are used to determine the complex impedance at that frequency. This is repeated for a range of frequencies to determine how the complex impedance of the system changes with stimulus frequency. Results are often displayed in a nyquist plot, showing the resistive and reactive components of impedance at each frequency measured. Often these plots will have characteristic shapes; for example, figure 1 shows the characteristic curved nyquist plot for a simple RC circuit. Figure 1: Nyquist Plot for a Simple RC Circuit[2] The ability to identify equivalent circuit components when analyzing bulk material properties can help the investigator identify physical properties of the material. For example, biological tissues have distinct impedance properties arising from their relative water content, cell packing density, and several other factors. Figure 2 shows an equivalent circuit model develop by Williams et. al. to explain the physiological significance of impedance properties recorded from subdurally implanted neuroelectrodes.

3 Figure 2: Equivalent Circuit Model of Neuroelectrode-Tissue Interface Impedance and Tissue Impedance[3] The aim of this study is to determine if pattern classification of electrical impedance spectroscopy measurements of breast tissue can be used as a viable means of pre-screening for cancerous tissue and to determine if any relevant physiological properties can be determined based on the strength of different extracted features. 2. Work Performed 2.1 Data Set The EIS data analyzed here was taken from a study by of 106 freshly excised breast tissue samples. EIS was performed in the range of 488 Hz to 1MHz and histology was performed to classify each tissue sample as either carcinoma, fibro-adenoma, mastopathy, glandular, connective, or adipose tissue. [4] This data is available publicly at the UCI Machine Learning Repository. [5] 2.2 Feature Extraction The publicly available data for this study does not include the raw EIS data recorded for each tissue sample. Rather, it contains a nine-attribute feature vector for each tissue sample, consisting of values extracted from the original data. The extracted attributes are as follows: I0: Impedance (ohm) at 0Hz PA500: Phase angle at 500kHz HFS: High-frequency slope of the phase angle DA: Distance between ends of the impedance spectrum AREA: Area under the impedance spectrum A/DA: AREA normalized by DA MAX IP: Maximum of the spectrum DR: Distance between I0 and the real component of the maximum frequency point P: Length of the spectral curve

4 2.3 Feature Selection There are two strategies that can be employed when selecting which features to use for classification: 1) rank each feature by its strength and select some subset of the best features 2) rank each unique combination of features by its strength and select the best combination of features. The first method is preferable in cases where the number of features is sufficiently large to make the task of finding the optimal feature set computationally difficult. The second method is favorable in cases where there a smaller number of features because it allows for the detection of interactions between features that improve classification results. Since there are only 9 features, there are only 511 possible feature vectors. Since this is a relatively low number, I performed an exhaustive search to determine the optimal feature vector from all possible feature vectors. To evaluate the performance of each potential feature vector, I randomly partitioned the data set into approximately equal sized testing and training sets. An extreme learning machine (ELM) with 200 neurons in the hidden layer was trained using the prospective feature vector of the training data, then the output of the test data was computed and the classification rate determined (justification of the use of ELM is discussed in 2.4 below). For each potential feature vector, the data was repartitioned and classification rate was determined 1000 times to ensure the determined performance of each feature vector was not over or under-estimated by a 'lucky' or 'unlucky' partition. 2.4 Classification Figure 3: ELM-SVM Classification Method My first step in classifying the data was to recreate a classification strategy developed by MR Daliri. [6] The chosen feature set is classified by a set of three ELMs with differing numbers of hidden nodes. An ELM is a single-layer perception in which the weights and offsets of connections between the input layer and hidden layer are fixed and randomly assigned. This greatly reduces the training time because only the weights of the output layer need to be computed. This is achieved by solving:

5 [6] Where H is the output of the hidden layer, Γ is the weights of the output layer and L is the true class labels. The solution to this problem is: [6] Where H is the generalized inverse of H. This results in a single-step training of the ELM, which greatly improves speed at the cost of reduced accuracy. The ELM layer of this classification method uses three ELMs with 150, 170, and 200 neurons in the hidden layer. The outputs of these ELMs form a feature vector that is classified by a multi-class support vector machine (SVM). Two common ways to extend SVM use to more than two classes are the 1 vs. 1 method and the 1 vs. all method. In the 1 vs. 1 method, binary classification of every combination of two classes is computed for the test data and the outputs are added to a 'voting' matrix where the class with the most vote for each input is chosen as the output class. In the 1 vs. all method, each class is compared against all other classes and the class for which the output function is greatest is chosen for each input. I have chosen to use the 1 vs. 1 method for this project since there are only 6 classes, making this computationally feasible. To determine the classification rate of the ELM-SVM method, I used 3-fold cross validation. The training set was used to train the ELM layer, then the training output of the ELM layer was used to train the multi-class SVM. A simple linear kernel SVM was used to avoid overfitting. The test data was then applied to the ELM layer, then the test output of the ELM layer applied to the SVM. The SVM output was used to calculate classification rate. 3-fold cross validation was repeated 1000 times with a the partition being randomly selected for each repetition; this was done to ensure that the mean classification rate reflected the method of classification rather than an ideal or nonideal partitioning. This entire process was repeated for several of the top-rated feature sets as determined in the feature selection step. Additionally, I performed classification of the data using only the multi-class SVM with the ELM layer omitted. This followed the same 3-fold cross validation with 1000 replicates technique described above. Last, I performed two-class classification using the above techniques (substituting a single linear SVM for the multiclass SVM) where the first class consisted of carcinoma, fibro-adenoma, and mastopathy tissue samples and the second class consisted of glandular, connective, and adipose tissue samples. My reason for doing so is that the first three tissues are all cancerous tissue types while the second three are noncancerous.

6 3. Results 3.1 Feature Selection C l a s s i f i c a t i o n R a t e I O P D A D R A R E A P A P P I O P A # N e u r o n s i n H i d d e n L a y e r Figure 4: Classification Rate as a Function of Neurons in the Hidden Layer for Highest-ranked Feature Sets After determining the strength of every feature vector set using the technique described above, I selected the top three feature vector sets: [PA500, P], [P], and [I0, PA500]. Figure 4 shows the classification rates of these three feature vector sets as a function of number of neurons in the hidden layer for a single ELM. I have included the feature set [I0, P, DA, DR, AREA] because this was determined to be the optimal feature set in a previous study of this data set using analysis of mutual information between individual features and the true class labels. The three best feature vectors determined using the exhaustive search method clearly outperform that determined by the mutual information method.

7 3.2 Classification Table 1: Summary of results of the 6-class problem Feature Vector [I0 P DA DR AREA][PA500 P][I0 PA500] [P] All 9 Features Mean CR(ELM/SVM) Mean CR(SVM) CR Std Dev (ELM/SVM) CR Std Dev (SVM) Mean Time (ELM/SVM) Mean Time (SVM) Time std (ELM/SVM) Time std(svm) The SVM-only classification method outperformed the ELM-SVM method for all feature sets. Additionally, the SVM-only classification method was faster than the ELM-SVM method for all feature vectors except for the one containing all features. While the SVM-only method took longer to compute for this feature vector, it yielded the highest classification rate of the 6-class problem. Table 2: Summary of results of the 2-class problem Feature Vector [I0 P DA DR AREA][PA500 P][I0 PA500] [P] All 9 Features Mean CR(ELM/SVM) Mean CR(SVM) CR Std Dev (ELM/SVM) CR Std Dev (SVM) Mean Time (ELM/SVM) Mean Time (SVM) Time std (ELM/SVM) Time std(svm) Both classification techniques show improvement in classification rate for the 2-class problem compared to the 6-class problem. Again, the SVM-only technique outperforms the ELM-SVM technique for all feature vectors. Additionally, the SVM-only technique requires a much shorter computation time compared to the ELM-SVM technique. 4 Discussion Performance of the feature vectors selected by exhaustive search outperformed the previously selected feature vector in the ELM-SVM method, but all feature vectors showed similar performance for the SVM-only approach. The justification for using a non-exhaustive feature selection technique is

8 typically to save time. In a clinical application of breast tissue classification using EIS, the feature selection step would be performed when developing the system for classification rather than in-clinic. As such, the amount of time it takes to perform feature selection is unimportant as long as it can be achieved in a finite amount of time. When performing the exhaustive feature selection for 9 features, it took less than 10 minutes to determine the ideal feature vectors; the improvement in classification performance more than made up for this minor inconvenience in computation time. If classification of breast tissue using EIS is ever used as a clinical pre-screening process, it is my recommendation that an exhaustive search of possible feature vectors be used, so long as the set of possible features remains relatively small. The ELM-SVM method took longer on average to compute and had a lower classification rate than the SVM-only method. The low classification rate of the ELM layer resulted in a loss of information and gain in computation time for the ELM-SVM method. Based on these results, it seems that a SVM-only approach would be preferable to the ELM-SVM approach for clinical classification of breast tissue. Based on this data set, the ability to determine exact tissue type does not have a high enough classification rate to replace the need for histological and chemical screening of tissue samples. However, results of the two-class approach indicate that classification using EIS may be useful as a pre-screening process to determine if an excised sample is appropriate for further screening. This would reduce the chance that a sample sent for further screening results in an inconclusive result. While the classification rate is only ~80% for the two-class problem; this is based on a relatively small number of tissue samples. If a large-scale study of EIS properties of breast tissue were performed, it is possible that classification rate could be greatly improved. One of the aims of this study was to determine if there is any physiological significance to the strength of different features. In the two-class problem, the inclusion of I0 caused considerable improvement in the ELM-SVM classification. I0 is a measure of the real impedance (non-capacitive); it is possible that cancerous tissue and regular tissue differ in water content and/or ionic strength and thus have different real impedance properties. Future work could improve the accuracy of classification and our understanding of the physiological basis of tissue impedance properties by including equivalent circuit component values in the feature vector. The potential benefit of immediate tissue classification, or at least improved pre-screening of tissue could reduce the cost of screening and the amount of time a patient needs to wait before receiving their results. Before this technique can be used clinically, the classification rate must be greatly improved to prevent the occurrence of false positives.

9 References [1] American Cancer Society. [ ndcytologyspecimensforcancer/testing-biopsy-and-cytology-specimens-for-cancer-how-long-doestesting-take] Accessed 15 Dec [2] Lvovich, V. F. (2012). Impedance spectroscopy: applications to electrochemical and dielectric phenomena. John Wiley & Sons. [3] Williams, J. C., Hippensteel, J. A., Dilgen, J., Shain, W., & Kipke, D. R. (2007). Complex impedance spectroscopy for monitoring tissue responses to inserted neural implants. Journal of neural engineering, 4(4), 410. [4] Jossinet, J. (1996). Variability of impedivity in normal and pathological breast tissue. Medical and Biological Engineering and Computing, 34(5), [5] UCI Machine Learning Repository. [ Accessed 7 Dec [6] Daliri, M. R. (2013). Combining extreme learning machines using support vector machines for breast tissue classification. Computer methods in biomechanics and biomedical engineering, (ahead-ofprint), 1-7.

ANN Based Fault Classifier and Fault Locator for Double Circuit Transmission Line

ANN Based Fault Classifier and Fault Locator for Double Circuit Transmission Line International Journal of Computer Sciences and Engineering Open Access Research Paper Volume-4, Special Issue-2, April 2016 E-ISSN: 2347-2693 ANN Based Fault Classifier and Fault Locator for Double Circuit

More information

Music Mood Classification

Music Mood Classification Music Mood Classification CS 229 Project Report Jose Padial Ashish Goel Introduction The aim of the project was to develop a music mood classifier. There are many categories of mood into which songs may

More information

Doppler. Doppler. Doppler shift. Doppler Frequency. Doppler shift. Doppler shift. Chapter 19

Doppler. Doppler. Doppler shift. Doppler Frequency. Doppler shift. Doppler shift. Chapter 19 Doppler Doppler Chapter 19 A moving train with a trumpet player holding the same tone for a very long time travels from your left to your right. The tone changes relative the motion of you (receiver) and

More information

The Calculation of G rms

The Calculation of G rms The Calculation of G rms QualMark Corp. Neill Doertenbach The metric of G rms is typically used to specify and compare the energy in repetitive shock vibration systems. However, the method of arriving

More information

Electrochemical Impedance Spectroscopy (EIS): A Powerful and Cost- Effective Tool for Fuel Cell Diagnostics

Electrochemical Impedance Spectroscopy (EIS): A Powerful and Cost- Effective Tool for Fuel Cell Diagnostics Electrochemical Impedance Spectroscopy (EIS): A Powerful and Cost- Effective Tool for Fuel Cell Diagnostics Electrochemical Impedance Spectroscopy (EIS) is a powerful diagnostic tool that you can use to

More information

Analysis of kiva.com Microlending Service! Hoda Eydgahi Julia Ma Andy Bardagjy December 9, 2010 MAS.622j

Analysis of kiva.com Microlending Service! Hoda Eydgahi Julia Ma Andy Bardagjy December 9, 2010 MAS.622j Analysis of kiva.com Microlending Service! Hoda Eydgahi Julia Ma Andy Bardagjy December 9, 2010 MAS.622j What is Kiva? An organization that allows people to lend small amounts of money via the Internet

More information

Chapter 4. Probability and Probability Distributions

Chapter 4. Probability and Probability Distributions Chapter 4. robability and robability Distributions Importance of Knowing robability To know whether a sample is not identical to the population from which it was selected, it is necessary to assess the

More information

Predicting Flight Delays

Predicting Flight Delays Predicting Flight Delays Dieterich Lawson jdlawson@stanford.edu William Castillo will.castillo@stanford.edu Introduction Every year approximately 20% of airline flights are delayed or cancelled, costing

More information

Impedance Matching and Matching Networks. Valentin Todorow, December, 2009

Impedance Matching and Matching Networks. Valentin Todorow, December, 2009 Impedance Matching and Matching Networks Valentin Todorow, December, 2009 RF for Plasma Processing - Definition of RF What is RF? The IEEE Standard Dictionary of Electrical and Electronics Terms defines

More information

Monotonicity Hints. Abstract

Monotonicity Hints. Abstract Monotonicity Hints Joseph Sill Computation and Neural Systems program California Institute of Technology email: joe@cs.caltech.edu Yaser S. Abu-Mostafa EE and CS Deptartments California Institute of Technology

More information

The Phase Modulator In NBFM Voice Communication Systems

The Phase Modulator In NBFM Voice Communication Systems The Phase Modulator In NBFM Voice Communication Systems Virgil Leenerts 8 March 5 The phase modulator has been a point of discussion as to why it is used and not a frequency modulator in what are called

More information

Machine Learning Logistic Regression

Machine Learning Logistic Regression Machine Learning Logistic Regression Jeff Howbert Introduction to Machine Learning Winter 2012 1 Logistic regression Name is somewhat misleading. Really a technique for classification, not regression.

More information

Mathematical Models of Supervised Learning and their Application to Medical Diagnosis

Mathematical Models of Supervised Learning and their Application to Medical Diagnosis Genomic, Proteomic and Transcriptomic Lab High Performance Computing and Networking Institute National Research Council, Italy Mathematical Models of Supervised Learning and their Application to Medical

More information

Sophomore Physics Laboratory (PH005/105)

Sophomore Physics Laboratory (PH005/105) CALIFORNIA INSTITUTE OF TECHNOLOGY PHYSICS MATHEMATICS AND ASTRONOMY DIVISION Sophomore Physics Laboratory (PH5/15) Analog Electronics Active Filters Copyright c Virgínio de Oliveira Sannibale, 23 (Revision

More information

Power System review W I L L I A M V. T O R R E A P R I L 1 0, 2 0 1 3

Power System review W I L L I A M V. T O R R E A P R I L 1 0, 2 0 1 3 Power System review W I L L I A M V. T O R R E A P R I L 1 0, 2 0 1 3 Basics of Power systems Network topology Transmission and Distribution Load and Resource Balance Economic Dispatch Steady State System

More information

High Productivity Data Processing Analytics Methods with Applications

High Productivity Data Processing Analytics Methods with Applications High Productivity Data Processing Analytics Methods with Applications Dr. Ing. Morris Riedel et al. Adjunct Associate Professor School of Engineering and Natural Sciences, University of Iceland Research

More information

Comparison of Non-linear Dimensionality Reduction Techniques for Classification with Gene Expression Microarray Data

Comparison of Non-linear Dimensionality Reduction Techniques for Classification with Gene Expression Microarray Data CMPE 59H Comparison of Non-linear Dimensionality Reduction Techniques for Classification with Gene Expression Microarray Data Term Project Report Fatma Güney, Kübra Kalkan 1/15/2013 Keywords: Non-linear

More information

Frequency Response of Filters

Frequency Response of Filters School of Engineering Department of Electrical and Computer Engineering 332:224 Principles of Electrical Engineering II Laboratory Experiment 2 Frequency Response of Filters 1 Introduction Objectives To

More information

THE PER-UNIT SYSTEM. (2) The per-unit values for various components lie within a narrow range regardless of the equipment rating.

THE PER-UNIT SYSTEM. (2) The per-unit values for various components lie within a narrow range regardless of the equipment rating. THE PER-UNIT SYSTEM An interconnected power system typically consists of many different voltage levels given a system containing several transformers and/or rotating machines. The per-unit system simplifies

More information

Inductors in AC Circuits

Inductors in AC Circuits Inductors in AC Circuits Name Section Resistors, inductors, and capacitors all have the effect of modifying the size of the current in an AC circuit and the time at which the current reaches its maximum

More information

Equipment: Power Supply, DAI, Transformer (8341), Variable resistance (8311), Variable inductance (8321), Variable capacitance (8331)

Equipment: Power Supply, DAI, Transformer (8341), Variable resistance (8311), Variable inductance (8321), Variable capacitance (8331) Lab 5: Single-phase transformer operations. Objective: to examine the design of single-phase transformers; to study the voltage and current ratios of transformers; to study the voltage regulation of the

More information

Employer Health Insurance Premium Prediction Elliott Lui

Employer Health Insurance Premium Prediction Elliott Lui Employer Health Insurance Premium Prediction Elliott Lui 1 Introduction The US spends 15.2% of its GDP on health care, more than any other country, and the cost of health insurance is rising faster than

More information

Data Mining. Nonlinear Classification

Data Mining. Nonlinear Classification Data Mining Unit # 6 Sajjad Haider Fall 2014 1 Nonlinear Classification Classes may not be separable by a linear boundary Suppose we randomly generate a data set as follows: X has range between 0 to 15

More information

LAB 7 MOSFET CHARACTERISTICS AND APPLICATIONS

LAB 7 MOSFET CHARACTERISTICS AND APPLICATIONS LAB 7 MOSFET CHARACTERISTICS AND APPLICATIONS Objective In this experiment you will study the i-v characteristics of an MOS transistor. You will use the MOSFET as a variable resistor and as a switch. BACKGROUND

More information

USB 3.0 CDR Model White Paper Revision 0.5

USB 3.0 CDR Model White Paper Revision 0.5 USB 3.0 CDR Model White Paper Revision 0.5 January 15, 2009 INTELLECTUAL PROPERTY DISCLAIMER THIS WHITE PAPER IS PROVIDED TO YOU AS IS WITH NO WARRANTIES WHATSOEVER, INCLUDING ANY WARRANTY OF MERCHANTABILITY,

More information

RLC Resonant Circuits

RLC Resonant Circuits C esonant Circuits Andrew McHutchon April 20, 203 Capacitors and Inductors There is a lot of inconsistency when it comes to dealing with reactances of complex components. The format followed in this document

More information

EDEXCEL NATIONAL CERTIFICATE/DIPLOMA UNIT 5 - ELECTRICAL AND ELECTRONIC PRINCIPLES NQF LEVEL 3 OUTCOME 4 - ALTERNATING CURRENT

EDEXCEL NATIONAL CERTIFICATE/DIPLOMA UNIT 5 - ELECTRICAL AND ELECTRONIC PRINCIPLES NQF LEVEL 3 OUTCOME 4 - ALTERNATING CURRENT EDEXCEL NATIONAL CERTIFICATE/DIPLOMA UNIT 5 - ELECTRICAL AND ELECTRONIC PRINCIPLES NQF LEVEL 3 OUTCOME 4 - ALTERNATING CURRENT 4 Understand single-phase alternating current (ac) theory Single phase AC

More information

Beating the NCAA Football Point Spread

Beating the NCAA Football Point Spread Beating the NCAA Football Point Spread Brian Liu Mathematical & Computational Sciences Stanford University Patrick Lai Computer Science Department Stanford University December 10, 2010 1 Introduction Over

More information

An Introduction to Electrochemical Impedance Spectroscopy (EIS)

An Introduction to Electrochemical Impedance Spectroscopy (EIS) An Introduction to Electrochemical Impedance Spectroscopy (EIS) C. Reece Excellent references: Identification of Electrochemical Processes by Frequency Response Analysis http://www.solartronanalytical.com/technicalsupport/technicalnotes/technote04.htm

More information

A wave lab inside a coaxial cable

A wave lab inside a coaxial cable INSTITUTE OF PHYSICS PUBLISHING Eur. J. Phys. 25 (2004) 581 591 EUROPEAN JOURNAL OF PHYSICS PII: S0143-0807(04)76273-X A wave lab inside a coaxial cable JoãoMSerra,MiguelCBrito,JMaiaAlves and A M Vallera

More information

Electroencephalography Analysis Using Neural Network and Support Vector Machine during Sleep

Electroencephalography Analysis Using Neural Network and Support Vector Machine during Sleep Engineering, 23, 5, 88-92 doi:.4236/eng.23.55b8 Published Online May 23 (http://www.scirp.org/journal/eng) Electroencephalography Analysis Using Neural Network and Support Vector Machine during Sleep JeeEun

More information

Step Response of RC Circuits

Step Response of RC Circuits Step Response of RC Circuits 1. OBJECTIVES...2 2. REFERENCE...2 3. CIRCUITS...2 4. COMPONENTS AND SPECIFICATIONS...3 QUANTITY...3 DESCRIPTION...3 COMMENTS...3 5. DISCUSSION...3 5.1 SOURCE RESISTANCE...3

More information

HYBRID PROBABILITY BASED ENSEMBLES FOR BANKRUPTCY PREDICTION

HYBRID PROBABILITY BASED ENSEMBLES FOR BANKRUPTCY PREDICTION HYBRID PROBABILITY BASED ENSEMBLES FOR BANKRUPTCY PREDICTION Chihli Hung 1, Jing Hong Chen 2, Stefan Wermter 3, 1,2 Department of Management Information Systems, Chung Yuan Christian University, Taiwan

More information

Understanding Power Impedance Supply for Optimum Decoupling

Understanding Power Impedance Supply for Optimum Decoupling Introduction Noise in power supplies is not only caused by the power supply itself, but also the load s interaction with the power supply (i.e. dynamic loads, switching, etc.). To lower load induced noise,

More information

Artificial Neural Networks and Support Vector Machines. CS 486/686: Introduction to Artificial Intelligence

Artificial Neural Networks and Support Vector Machines. CS 486/686: Introduction to Artificial Intelligence Artificial Neural Networks and Support Vector Machines CS 486/686: Introduction to Artificial Intelligence 1 Outline What is a Neural Network? - Perceptron learners - Multi-layer networks What is a Support

More information

Data Mining Techniques for Prognosis in Pancreatic Cancer

Data Mining Techniques for Prognosis in Pancreatic Cancer Data Mining Techniques for Prognosis in Pancreatic Cancer by Stuart Floyd A Thesis Submitted to the Faculty of the WORCESTER POLYTECHNIC INSTITUE In partial fulfillment of the requirements for the Degree

More information

Journal of Information

Journal of Information Journal of Information journal homepage: http://www.pakinsight.com/?ic=journal&journal=104 PREDICT SURVIVAL OF PATIENTS WITH LUNG CANCER USING AN ENSEMBLE FEATURE SELECTION ALGORITHM AND CLASSIFICATION

More information

Jitter Measurements in Serial Data Signals

Jitter Measurements in Serial Data Signals Jitter Measurements in Serial Data Signals Michael Schnecker, Product Manager LeCroy Corporation Introduction The increasing speed of serial data transmission systems places greater importance on measuring

More information

AN-837 APPLICATION NOTE

AN-837 APPLICATION NOTE APPLICATION NOTE One Technology Way P.O. Box 916 Norwood, MA 262-916, U.S.A. Tel: 781.329.47 Fax: 781.461.3113 www.analog.com DDS-Based Clock Jitter Performance vs. DAC Reconstruction Filter Performance

More information

An Introduction to Electrochemical Impedance Measurement

An Introduction to Electrochemical Impedance Measurement An Introduction to Electrochemical Impedance Measurement Technical Report No. 6 An Introduction to Electrochemical Impedance Measurement N D Cogger and N J Evans Technical Report No. 6 Part No.: BTR006

More information

Case Studies in On-Line Measurement of PD in Motors Fed by Voltage Source PWM Drives

Case Studies in On-Line Measurement of PD in Motors Fed by Voltage Source PWM Drives Case Studies in On-Line Measurement of PD in Motors Fed by Voltage Source PWM Drives G.C. Stone, I. Culbert, H.G. Sedding Qualitrol-Iris Power Mississauga, Ontario, Canada Abstract On-line partial discharge

More information

Time Domain and Frequency Domain Techniques For Multi Shaker Time Waveform Replication

Time Domain and Frequency Domain Techniques For Multi Shaker Time Waveform Replication Time Domain and Frequency Domain Techniques For Multi Shaker Time Waveform Replication Thomas Reilly Data Physics Corporation 1741 Technology Drive, Suite 260 San Jose, CA 95110 (408) 216-8440 This paper

More information

Artificial Neural Network, Decision Tree and Statistical Techniques Applied for Designing and Developing E-mail Classifier

Artificial Neural Network, Decision Tree and Statistical Techniques Applied for Designing and Developing E-mail Classifier International Journal of Recent Technology and Engineering (IJRTE) ISSN: 2277-3878, Volume-1, Issue-6, January 2013 Artificial Neural Network, Decision Tree and Statistical Techniques Applied for Designing

More information

ANN Based Modeling of High Speed IC Interconnects. Q.J. Zhang, Carleton University

ANN Based Modeling of High Speed IC Interconnects. Q.J. Zhang, Carleton University ANN Based Modeling of High Speed IC Interconnects Needs for Repeated Simulation Signal integrity optimization Iterative design and re-optimization Monte-Carlo analysis Yield optimization Iterative design

More information

SIGNAL GENERATORS and OSCILLOSCOPE CALIBRATION

SIGNAL GENERATORS and OSCILLOSCOPE CALIBRATION 1 SIGNAL GENERATORS and OSCILLOSCOPE CALIBRATION By Lannes S. Purnell FLUKE CORPORATION 2 This paper shows how standard signal generators can be used as leveled sine wave sources for calibrating oscilloscopes.

More information

Active Learning SVM for Blogs recommendation

Active Learning SVM for Blogs recommendation Active Learning SVM for Blogs recommendation Xin Guan Computer Science, George Mason University Ⅰ.Introduction In the DH Now website, they try to review a big amount of blogs and articles and find the

More information

Author Gender Identification of English Novels

Author Gender Identification of English Novels Author Gender Identification of English Novels Joseph Baena and Catherine Chen December 13, 2013 1 Introduction Machine learning algorithms have long been used in studies of authorship, particularly in

More information

Use and Application of Output Limiting Amplifiers (HFA1115, HFA1130, HFA1135)

Use and Application of Output Limiting Amplifiers (HFA1115, HFA1130, HFA1135) Use and Application of Output Limiting Amplifiers (HFA111, HFA110, HFA11) Application Note November 1996 AN96 Introduction Amplifiers with internal voltage clamps, also known as limiting amplifiers, have

More information

Series and Parallel Resistive Circuits

Series and Parallel Resistive Circuits Series and Parallel Resistive Circuits The configuration of circuit elements clearly affects the behaviour of a circuit. Resistors connected in series or in parallel are very common in a circuit and act

More information

Chapter 2 The Research on Fault Diagnosis of Building Electrical System Based on RBF Neural Network

Chapter 2 The Research on Fault Diagnosis of Building Electrical System Based on RBF Neural Network Chapter 2 The Research on Fault Diagnosis of Building Electrical System Based on RBF Neural Network Qian Wu, Yahui Wang, Long Zhang and Li Shen Abstract Building electrical system fault diagnosis is the

More information

Functional Data Analysis of MALDI TOF Protein Spectra

Functional Data Analysis of MALDI TOF Protein Spectra Functional Data Analysis of MALDI TOF Protein Spectra Dean Billheimer dean.billheimer@vanderbilt.edu. Department of Biostatistics Vanderbilt University Vanderbilt Ingram Cancer Center FDA for MALDI TOF

More information

Introduction to Digital Audio

Introduction to Digital Audio Introduction to Digital Audio Before the development of high-speed, low-cost digital computers and analog-to-digital conversion circuits, all recording and manipulation of sound was done using analog techniques.

More information

LR Phono Preamps. Pete Millett ETF.13. pmillett@hotmail.com

LR Phono Preamps. Pete Millett ETF.13. pmillett@hotmail.com LR Phono Preamps Pete Millett ETF.13 pmillett@hotmail.com Agenda A bit about me Part 1: What is, and why use, RIAA? Grooves on records The RIAA standard Implementations of RIAA EQ networks and preamps

More information

Scalable Developments for Big Data Analytics in Remote Sensing

Scalable Developments for Big Data Analytics in Remote Sensing Scalable Developments for Big Data Analytics in Remote Sensing Federated Systems and Data Division Research Group High Productivity Data Processing Dr.-Ing. Morris Riedel et al. Research Group Leader,

More information

Predict Influencers in the Social Network

Predict Influencers in the Social Network Predict Influencers in the Social Network Ruishan Liu, Yang Zhao and Liuyu Zhou Email: rliu2, yzhao2, lyzhou@stanford.edu Department of Electrical Engineering, Stanford University Abstract Given two persons

More information

Big Data Analytics CSCI 4030

Big Data Analytics CSCI 4030 High dim. data Graph data Infinite data Machine learning Apps Locality sensitive hashing PageRank, SimRank Filtering data streams SVM Recommen der systems Clustering Community Detection Web advertising

More information

Neural Networks and Support Vector Machines

Neural Networks and Support Vector Machines INF5390 - Kunstig intelligens Neural Networks and Support Vector Machines Roar Fjellheim INF5390-13 Neural Networks and SVM 1 Outline Neural networks Perceptrons Neural networks Support vector machines

More information

OplAnalyzer: A Toolbox for MALDI-TOF Mass Spectrometry Data Analysis

OplAnalyzer: A Toolbox for MALDI-TOF Mass Spectrometry Data Analysis OplAnalyzer: A Toolbox for MALDI-TOF Mass Spectrometry Data Analysis Thang V. Pham and Connie R. Jimenez OncoProteomics Laboratory, Cancer Center Amsterdam, VU University Medical Center De Boelelaan 1117,

More information

Laboratory 4: Feedback and Compensation

Laboratory 4: Feedback and Compensation Laboratory 4: Feedback and Compensation To be performed during Week 9 (Oct. 20-24) and Week 10 (Oct. 27-31) Due Week 11 (Nov. 3-7) 1 Pre-Lab This Pre-Lab should be completed before attending your regular

More information

Title : Analog Circuit for Sound Localization Applications

Title : Analog Circuit for Sound Localization Applications Title : Analog Circuit for Sound Localization Applications Author s Name : Saurabh Kumar Tiwary Brett Diamond Andrea Okerholm Contact Author : Saurabh Kumar Tiwary A-51 Amberson Plaza 5030 Center Avenue

More information

TENSILE TESTING PRACTICAL

TENSILE TESTING PRACTICAL TENSILE TESTING PRACTICAL MTK 2B- Science Of Materials Ts epo Mputsoe 215024596 Summary Material have different properties all varying form mechanical to chemical properties. Taking special interest in

More information

Towards better accuracy for Spam predictions

Towards better accuracy for Spam predictions Towards better accuracy for Spam predictions Chengyan Zhao Department of Computer Science University of Toronto Toronto, Ontario, Canada M5S 2E4 czhao@cs.toronto.edu Abstract Spam identification is crucial

More information

Nodal and Loop Analysis

Nodal and Loop Analysis Nodal and Loop Analysis The process of analyzing circuits can sometimes be a difficult task to do. Examining a circuit with the node or loop methods can reduce the amount of time required to get important

More information

W03 Analysis of DC Circuits. Yrd. Doç. Dr. Aytaç Gören

W03 Analysis of DC Circuits. Yrd. Doç. Dr. Aytaç Gören W03 Analysis of DC Circuits Yrd. Doç. Dr. Aytaç Gören ELK 2018 - Contents W01 Basic Concepts in Electronics W02 AC to DC Conversion W03 Analysis of DC Circuits (self and condenser) W04 Transistors and

More information

Comparing the Results of Support Vector Machines with Traditional Data Mining Algorithms

Comparing the Results of Support Vector Machines with Traditional Data Mining Algorithms Comparing the Results of Support Vector Machines with Traditional Data Mining Algorithms Scott Pion and Lutz Hamel Abstract This paper presents the results of a series of analyses performed on direct mail

More information

Evaluation of Machine Learning Techniques for Green Energy Prediction

Evaluation of Machine Learning Techniques for Green Energy Prediction arxiv:1406.3726v1 [cs.lg] 14 Jun 2014 Evaluation of Machine Learning Techniques for Green Energy Prediction 1 Objective Ankur Sahai University of Mainz, Germany We evaluate Machine Learning techniques

More information

Objectives. Experimentally determine the yield strength, tensile strength, and modules of elasticity and ductility of given materials.

Objectives. Experimentally determine the yield strength, tensile strength, and modules of elasticity and ductility of given materials. Lab 3 Tension Test Objectives Concepts Background Experimental Procedure Report Requirements Discussion Objectives Experimentally determine the yield strength, tensile strength, and modules of elasticity

More information

Lecture 6: Classification & Localization. boris. ginzburg@intel.com

Lecture 6: Classification & Localization. boris. ginzburg@intel.com Lecture 6: Classification & Localization boris. ginzburg@intel.com 1 Agenda ILSVRC 2014 Overfeat: integrated classification, localization, and detection Classification with Localization Detection. 2 ILSVRC-2014

More information

Spike-Based Sensing and Processing: What are spikes good for? John G. Harris Electrical and Computer Engineering Dept

Spike-Based Sensing and Processing: What are spikes good for? John G. Harris Electrical and Computer Engineering Dept Spike-Based Sensing and Processing: What are spikes good for? John G. Harris Electrical and Computer Engineering Dept ONR NEURO-SILICON WORKSHOP, AUG 1-2, 2006 Take Home Messages Introduce integrate-and-fire

More information

BASIC ELECTRONICS AC CIRCUIT ANALYSIS. December 2011

BASIC ELECTRONICS AC CIRCUIT ANALYSIS. December 2011 AM 5-202 BASIC ELECTRONICS AC CIRCUIT ANALYSIS December 2011 DISTRIBUTION RESTRICTION: Approved for Pubic Release. Distribution is unlimited. DEPARTMENT OF THE ARMY MILITARY AUXILIARY RADIO SYSTEM FORT

More information

FRAUD DETECTION IN ELECTRIC POWER DISTRIBUTION NETWORKS USING AN ANN-BASED KNOWLEDGE-DISCOVERY PROCESS

FRAUD DETECTION IN ELECTRIC POWER DISTRIBUTION NETWORKS USING AN ANN-BASED KNOWLEDGE-DISCOVERY PROCESS FRAUD DETECTION IN ELECTRIC POWER DISTRIBUTION NETWORKS USING AN ANN-BASED KNOWLEDGE-DISCOVERY PROCESS Breno C. Costa, Bruno. L. A. Alberto, André M. Portela, W. Maduro, Esdras O. Eler PDITec, Belo Horizonte,

More information

Classification algorithm in Data mining: An Overview

Classification algorithm in Data mining: An Overview Classification algorithm in Data mining: An Overview S.Neelamegam #1, Dr.E.Ramaraj *2 #1 M.phil Scholar, Department of Computer Science and Engineering, Alagappa University, Karaikudi. *2 Professor, Department

More information

Feature Selection using Integer and Binary coded Genetic Algorithm to improve the performance of SVM Classifier

Feature Selection using Integer and Binary coded Genetic Algorithm to improve the performance of SVM Classifier Feature Selection using Integer and Binary coded Genetic Algorithm to improve the performance of SVM Classifier D.Nithya a, *, V.Suganya b,1, R.Saranya Irudaya Mary c,1 Abstract - This paper presents,

More information

Solving Mass Balances using Matrix Algebra

Solving Mass Balances using Matrix Algebra Page: 1 Alex Doll, P.Eng, Alex G Doll Consulting Ltd. http://www.agdconsulting.ca Abstract Matrix Algebra, also known as linear algebra, is well suited to solving material balance problems encountered

More information

AVX EMI SOLUTIONS Ron Demcko, Fellow of AVX Corporation Chris Mello, Principal Engineer, AVX Corporation Brian Ward, Business Manager, AVX Corporation

AVX EMI SOLUTIONS Ron Demcko, Fellow of AVX Corporation Chris Mello, Principal Engineer, AVX Corporation Brian Ward, Business Manager, AVX Corporation AVX EMI SOLUTIONS Ron Demcko, Fellow of AVX Corporation Chris Mello, Principal Engineer, AVX Corporation Brian Ward, Business Manager, AVX Corporation Abstract EMC compatibility is becoming a key design

More information

Application Note. So You Need to Measure Some Inductors?

Application Note. So You Need to Measure Some Inductors? So You Need to Measure Some nductors? Take a look at the 1910 nductance Analyzer. Although specifically designed for production testing of inductors and coils, in addition to measuring inductance (L),

More information

RLC Series Resonance

RLC Series Resonance RLC Series Resonance 11EM Object: The purpose of this laboratory activity is to study resonance in a resistor-inductor-capacitor (RLC) circuit by examining the current through the circuit as a function

More information

KNOWLEDGE BASED ANALYSIS OF VARIOUS STATISTICAL TOOLS IN DETECTING BREAST CANCER

KNOWLEDGE BASED ANALYSIS OF VARIOUS STATISTICAL TOOLS IN DETECTING BREAST CANCER KNOWLEDGE BASED ANALYSIS OF VARIOUS STATISTICAL TOOLS IN DETECTING BREAST CANCER S. Aruna 1, Dr S.P. Rajagopalan 2 and L.V. Nandakishore 3 1,2 Department of Computer Applications, Dr M.G.R University,

More information

Author: Dr. Society of Electrophysio. Reference: Electrodes. should include: electrode shape size use. direction.

Author: Dr. Society of Electrophysio. Reference: Electrodes. should include: electrode shape size use. direction. Standards for Reportin ng EMG Data Author: Dr. Roberto Merletti, Politecnico di Torino, Italy The Standards for Reporting EMG Data, written by Dr. Robertoo Merletti, are endorsed by the International Society

More information

Programmable-Gain Transimpedance Amplifiers Maximize Dynamic Range in Spectroscopy Systems

Programmable-Gain Transimpedance Amplifiers Maximize Dynamic Range in Spectroscopy Systems Programmable-Gain Transimpedance Amplifiers Maximize Dynamic Range in Spectroscopy Systems PHOTODIODE VOLTAGE SHORT-CIRCUIT PHOTODIODE SHORT- CIRCUIT VOLTAGE 0mV DARK ark By Luis Orozco Introduction Precision

More information

RF Measurements Using a Modular Digitizer

RF Measurements Using a Modular Digitizer RF Measurements Using a Modular Digitizer Modern modular digitizers, like the Spectrum M4i series PCIe digitizers, offer greater bandwidth and higher resolution at any given bandwidth than ever before.

More information

Server Load Prediction

Server Load Prediction Server Load Prediction Suthee Chaidaroon (unsuthee@stanford.edu) Joon Yeong Kim (kim64@stanford.edu) Jonghan Seo (jonghan@stanford.edu) Abstract Estimating server load average is one of the methods that

More information

Chapter 7. Feature Selection. 7.1 Introduction

Chapter 7. Feature Selection. 7.1 Introduction Chapter 7 Feature Selection Feature selection is not used in the system classification experiments, which will be discussed in Chapter 8 and 9. However, as an autonomous system, OMEGA includes feature

More information

Investigation the Protective Performance of Organic Coatings with Different Breakage Degree Using EIS United to SOM Neural Network

Investigation the Protective Performance of Organic Coatings with Different Breakage Degree Using EIS United to SOM Neural Network Int. J. Electrochem. Sci., 8 (213) 1895-192 International Journal of ELECTROCHEMICAL SCIENCE www.electrochemsci.org Investigation the Protective Performance of Organic Coatings with Different Breakage

More information

Cable Analysis and Fault Detection using the Bode 100

Cable Analysis and Fault Detection using the Bode 100 Cable Analysis and Fault Detection using the Bode 100 By Stephan Synkule 2014 by OMICRON Lab V1.3 Visit www.omicron-lab.com for more information. Contact support@omicron-lab.com for technical support.

More information

LOW COST MOTOR PROTECTION FILTERS FOR PWM DRIVE APPLICATIONS STOPS MOTOR DAMAGE

LOW COST MOTOR PROTECTION FILTERS FOR PWM DRIVE APPLICATIONS STOPS MOTOR DAMAGE LOW COST MOTOR PROTECTION FILTERS FOR PWM DRIVE APPLICATIONS STOPS MOTOR DAMAGE Karl M. Hink, Executive Vice President Originally presented at the Power Quality 99 Conference ABSTRACT Motor protection

More information

Novelty Detection in image recognition using IRF Neural Networks properties

Novelty Detection in image recognition using IRF Neural Networks properties Novelty Detection in image recognition using IRF Neural Networks properties Philippe Smagghe, Jean-Luc Buessler, Jean-Philippe Urban Université de Haute-Alsace MIPS 4, rue des Frères Lumière, 68093 Mulhouse,

More information

TCOM 370 NOTES 99-4 BANDWIDTH, FREQUENCY RESPONSE, AND CAPACITY OF COMMUNICATION LINKS

TCOM 370 NOTES 99-4 BANDWIDTH, FREQUENCY RESPONSE, AND CAPACITY OF COMMUNICATION LINKS TCOM 370 NOTES 99-4 BANDWIDTH, FREQUENCY RESPONSE, AND CAPACITY OF COMMUNICATION LINKS 1. Bandwidth: The bandwidth of a communication link, or in general any system, was loosely defined as the width of

More information

Sense Making in an IOT World: Sensor Data Analysis with Deep Learning

Sense Making in an IOT World: Sensor Data Analysis with Deep Learning Sense Making in an IOT World: Sensor Data Analysis with Deep Learning Natalia Vassilieva, PhD Senior Research Manager GTC 2016 Deep learning proof points as of today Vision Speech Text Other Search & information

More information

Data Mining Practical Machine Learning Tools and Techniques

Data Mining Practical Machine Learning Tools and Techniques Ensemble learning Data Mining Practical Machine Learning Tools and Techniques Slides for Chapter 8 of Data Mining by I. H. Witten, E. Frank and M. A. Hall Combining multiple models Bagging The basic idea

More information

Large Margin DAGs for Multiclass Classification

Large Margin DAGs for Multiclass Classification S.A. Solla, T.K. Leen and K.-R. Müller (eds.), 57 55, MIT Press (000) Large Margin DAGs for Multiclass Classification John C. Platt Microsoft Research Microsoft Way Redmond, WA 9805 jplatt@microsoft.com

More information

HD Radio FM Transmission System Specifications Rev. F August 24, 2011

HD Radio FM Transmission System Specifications Rev. F August 24, 2011 HD Radio FM Transmission System Specifications Rev. F August 24, 2011 SY_SSS_1026s TRADEMARKS HD Radio and the HD, HD Radio, and Arc logos are proprietary trademarks of ibiquity Digital Corporation. ibiquity,

More information

MAXIMIZING RETURN ON DIRECT MARKETING CAMPAIGNS

MAXIMIZING RETURN ON DIRECT MARKETING CAMPAIGNS MAXIMIZING RETURN ON DIRET MARKETING AMPAIGNS IN OMMERIAL BANKING S 229 Project: Final Report Oleksandra Onosova INTRODUTION Recent innovations in cloud computing and unified communications have made a

More information

PIEZO FILTERS INTRODUCTION

PIEZO FILTERS INTRODUCTION For more than two decades, ceramic filter technology has been instrumental in the proliferation of solid state electronics. A view of the future reveals that even greater expectations will be placed on

More information

Car Insurance. Havránek, Pokorný, Tomášek

Car Insurance. Havránek, Pokorný, Tomášek Car Insurance Havránek, Pokorný, Tomášek Outline Data overview Horizontal approach + Decision tree/forests Vertical (column) approach + Neural networks SVM Data overview Customers Viewed policies Bought

More information

30. Bode Plots. Introduction

30. Bode Plots. Introduction 0. Bode Plots Introduction Each of the circuits in this problem set is represented by a magnitude Bode plot. The network function provides a connection between the Bode plot and the circuit. To solve these

More information

Analog and Digital Signals, Time and Frequency Representation of Signals

Analog and Digital Signals, Time and Frequency Representation of Signals 1 Analog and Digital Signals, Time and Frequency Representation of Signals Required reading: Garcia 3.1, 3.2 CSE 3213, Fall 2010 Instructor: N. Vlajic 2 Data vs. Signal Analog vs. Digital Analog Signals

More information

PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 4: LINEAR MODELS FOR CLASSIFICATION

PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 4: LINEAR MODELS FOR CLASSIFICATION PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 4: LINEAR MODELS FOR CLASSIFICATION Introduction In the previous chapter, we explored a class of regression models having particularly simple analytical

More information

Frequency domain application of the Hot-Spot method for the fatigue assessment of the weld seams

Frequency domain application of the Hot-Spot method for the fatigue assessment of the weld seams Frequency domain application of the Hot-Spot method for the fatigue assessment of the weld seams Dr. Ing. Sauro Vannicola 1 sauro.vannicola@ch.abb.com Dr. Ing. Luigi De Mercato 2 luigi.de-mercato@ch.abb.com

More information

ENSEMBLE DECISION TREE CLASSIFIER FOR BREAST CANCER DATA

ENSEMBLE DECISION TREE CLASSIFIER FOR BREAST CANCER DATA ENSEMBLE DECISION TREE CLASSIFIER FOR BREAST CANCER DATA D.Lavanya 1 and Dr.K.Usha Rani 2 1 Research Scholar, Department of Computer Science, Sree Padmavathi Mahila Visvavidyalayam, Tirupati, Andhra Pradesh,

More information