Chapter IV: Discriminant Analysis

Tatsuoka, Maurice M.; Tiedeman, David V.

doi:10.3102/00346543024005402

Cited by 14 publications

(6 citation statements)

References 92 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…According to theories of DA, there is very little chance of DA classification to be different from the CA classification. In rare cases, when the DA classification is different from CA classification, then the CA classification is replaced with DA classification as evidence from literature which suggests that the DA classification can help remove the misclassification that often plagues the CA output [23].…”

Section: Discriminant Analysis (Da)mentioning

confidence: 99%

See 1 more Smart Citation

An empirical study of the effect of learning styles on the faults found during the software requirements inspection

Goswami

Walia

2013

2013 IEEE 24th International Symposium on Software Reliability Engineering (ISSRE)

View full text Add to dashboard Cite

Inspections aid software managers by early detection and removal of faults committed during the creation of requirements and design documents. This helps reduce the rework during the later stages of software development. While inspections are effective in practice, the evidence suggests that the effectiveness of inspectors varies widely. Cognitive psychologists have used Learning Style (LS) to show the improvement in student's score by considering their characteristic strength and preferences to acquire and process information. This concept of LS can cross over to software engineering as a means of increasing the inspection effectiveness. This paper investigates the effect of the LS of inspectors on fault detection abilities of inspection teams and individual inspectors. Using the inspection data with varying number of participants, we analyzed the effect of the LS of inspectors across various inspection team sizes on the inspection performance. We also analyzed the effect of LS categories on the individual inspection performance. The initial results show that the teams composed of inspectors with different LS preferences are more effective and efficient than the teams of inspectors who had similar LS's. The results also provide insights into the LS categories that favor requirements inspection.

show abstract

Section: Discriminant Analysis (Da)mentioning

confidence: 99%

“…This result of the DA is used to maximize the LS variations across different clusters, and minimize the LS variations within each cluster [17,23].…”

Section: Discriminant Analysis (Da)mentioning

confidence: 99%

An empirical study of the effect of learning styles on the faults found during the software requirements inspection

Goswami

Walia

2013

2013 IEEE 24th International Symposium on Software Reliability Engineering (ISSRE)

View full text Add to dashboard Cite

show abstract

“…Discriminant analysis is a classification method based on separating data by a line or hyper-plane, depending on the number of class labels, to maximize the discrimination (or minimize misclassification) between classes [67,68]. The lines or hyper-planes are represented by linear combinations of predictor variables (or features), also known as discriminant function(s).…”

Section: Discriminant Analysismentioning

confidence: 99%

A Framework for Meta-heuristic Parameter Performance Prediction Using Fitness Landscape Analysis and Machine Learning

McDevitt,

Harrison,

Ombuki-Berman

2023

2023 IEEE Congress on Evolutionary Computation (CEC)

View full text Add to dashboard Cite

The behaviour of an optimization algorithm when attempting to solve a problem depends on the values assigned to its control parameters. For an algorithm to obtain desirable performance, its control parameter values must be chosen based on the current problem. Despite being necessary for optimal performance, selecting appropriate control parameter values is time-consuming, computationally expensive, and challenging. As the quantity of control parameters increases, so does the time complexity associated with searching for practical values, which often overshadows addressing the problem at hand, limiting the efficiency of an algorithm. As primarily recognized by the no free lunch theorem, there is no one-size-fits-all to problem-solving; hence from understanding a problem, a tailored approach can substantially help solve it.To predict the performance of control parameter configurations in unseen environments, this thesis crafts an intelligent generalizable framework leveraging machine learning classification and quantitative characteristics about the problem in question. The proposed parameter performance classifier (PPC) framework is extensively explored by training 84 high-accuracy classifiers comprised of multiple sampling methods, fitness types, and binning strategies. Furthermore, the novel framework is utilized in constructing a new parameter-free particle swarm optimization (PSO) variant called PPC-PSO that effectively eliminates the computational cost of parameter tuning, yields competitive performance amongst other leading methodologies across 99 benchmark functions, and is highly accessible to researchers and practitioners. The success of PPC-PSO shows excellent promise for the applicability of the PPC framework in making many more robust parameter-free meta-heuristic algorithms in the future with incredible generalization capabilities.I would like to personally express my dearest thanks to the following people for their involvement in this thesis and for the knowledge/experience I gained throughout the process:• Dr. Kyle R. Harrison for his excellent mentoring, constant involvement, and supervision, and for being a catalyst for growth and learning over the course of my thesis.• Dr. Beatrice M. Ombuki-Berman for her magnificent guidance, years of believing in me, distinguished supervision, and the opportunity to learn, grow as a person and be a part of a wonderful research team.

show abstract

“…Given the high dimensionality of embedding vectors, however, the rst step in the embedding analysis is reducing the dimension of representations. Dimensionality reduction algorithms such as PCA [53], LDA [70], and tSNE [29] Latent Discriminant Axis Among the most popular dimensionality reduction algorithms, SEER uses Linear Discriminant Analysis (LDA), as it recognizes the class labels and maximizes the separation between classes during the dimensionality reduction. Figure 7 shows the intuition behind how LDA performs high-dimensionality reduction.…”

Section: Embedding Analysis Seer Relies On Visualization Techniquesmentioning

confidence: 99%

Perfect is the enemy of test oracle

Ibrahimzada

Varli

Tekinoglu

et al. 2022

Proceedings of the 30th ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Enginee

View full text Add to dashboard Cite

Automation of test oracles is one of the most challenging facets of software testing, but remains comparatively less addressed compared to automated test input generation. Test oracles rely on a ground-truth that can distinguish between the correct and buggy behavior to determine whether a test fails (detects a bug) or passes. What makes the oracle problem challenging and undecidable is the assumption that the ground-truth should know the exact expected, correct, or buggy behavior. However, we argue that one can still build an accurate oracle without knowing the exact correct or buggy behavior, but how these two might di er. This paper presents SEER, a learning-based approach that in the absence of test assertions or other types of oracle, can determine whether a unit test passes or fails on a given method under test (MUT). To build the groundtruth, SEER jointly embeds unit tests and the implementation of MUTs into a uni ed vector space, in such a way that the neural representation of tests are similar to that of MUTs they pass on them, but dissimilar to MUTs they fail on them. The classi er built on top of this vector representation serves as the oracle to generate "fail" labels, when test inputs detect a bug in MUT or "pass" labels, otherwise. Our extensive experiments on applying SEER to more than 5K unit tests from a diverse set of open-source Java projects show that the produced oracle is (1) e ective in predicting the fail or pass labels, achieving an overall accuracy, precision, recall, and F1 measure of 93%, 86%, 94%, and 90%, (2) generalizable, predicting the labels for the unit test of projects that were not in training or validation set with negligible performance drop, and (3) e cient, detecting the existence of bugs in only 6.5 milliseconds on average. Moreover, by interpreting the neural model and looking at it beyond a closed-box solution, we con rm that the oracle is valid, i.e., it predicts the labels through learning relevant features. CCS CONCEPTS• Software and its engineering → Software testing and debugging; • Computing methodologies → Neural networks.

show abstract

Chapter IV: Discriminant Analysis

Cited by 14 publications

References 92 publications

An empirical study of the effect of learning styles on the faults found during the software requirements inspection

An empirical study of the effect of learning styles on the faults found during the software requirements inspection

A Framework for Meta-heuristic Parameter Performance Prediction Using Fitness Landscape Analysis and Machine Learning

Perfect is the enemy of test oracle

Contact Info

Product

Resources

About