Classification in high-dimensional spectral data: Accuracy vs. interpretability vs. model size

Backhaus, Andreas; Seiffert, Udo

doi:10.1016/j.neucom.2013.09.048

Cited by 34 publications

(29 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, metric learning can be integrated efficiently into the classification scheme, and results from learning theory can be derived by referring to the resulting function class. This has been demonstrated in the context of learning vector quantization (LVQ), where metric learning opened the way towards efficient state-of-the-art results in various areas, including biomedical data analysis, robotic vision, and spectral analysis [4,19,1]. Because of the intuitive definition of models in terms of prototypical representatives, prototype-based methods like LVQ enjoy a wide popularity in application domains, particularly if human inspection and interaction are necessary, or life-long model adaptation is considered [29,20,18].…”

Section: Motivation and Related Workmentioning

confidence: 99%

“…Again, the introduction of any gaps is crucial for class-discrimination, so a minimum of the error surface is expected for settings where both costs λ A− and λ B− , become high. Figure 3 shows E RGLVQ for configurations (λ A− , λ B− ) in increasing steps of 0.1 over the interval [0,1]. The remaining third parameter in λ is fixed to the final value after training, in this case it is close to the small constant λ AB ≈ ǫ.…”

Section: Rglvq Error Function Surfacementioning

confidence: 99%

See 1 more Smart Citation

Metric learning for sequences in relational LVQ

et al. 2015

View full text Add to dashboard Cite

Metric learning constitutes a well-investigated field for vectorial data with successful applications, e.g. in computer vision, information retrieval, or bioinformatics. One particularly promising approach is offered by lowrank metric adaptation integrated into modern variants of learning vector quantization (LVQ). This technique is scalable with respect to both, data dimensionality and the number of data points, and it can be accompanied by strong guarantees of learning theory. Recent extensions of LVQ to general (dis-)similarity data have paved the way towards LVQ classifiers for non-vectorial, possibly discrete, structured objects such as sequences, which are addressed by classical alignment in bioinformatics applications. In this context, the choice of metric parameters plays a crucial role for the result, just as it does in the vectorial setting. In this contribution, we propose a metric learning scheme which allows for an autonomous learning of parameters (such as the underlying scoring matrix in sequence alignments) according to a given discriminative task in relational LVQ. Besides facilitating the often crucial and problematic choice of the scoring parameters in applications, this extension offers an increased interpretability of the results by pointing out structural invariances for the given task. *

show abstract

Section: Motivation and Related Workmentioning

confidence: 99%

Section: Rglvq Error Function Surfacementioning

confidence: 99%

Metric learning for sequences in relational LVQ

et al. 2015

View full text Add to dashboard Cite

show abstract

“…Equation 5 consists of univariate or multivariate analyses; Raman band-based, component analysis score-based, or cluster-based component mapping; clustering and classification procedures; mixture deconvolution; or other ways in which to selectively emphasize or quantify the information of interest. [35][36][37][38] We next define preprocessing and data analysis as stages while, within each stage, steps denote those particular processes required to achieve specific objectives (more details below). Eqs.…”

Section: Theorymentioning

confidence: 99%

Development and Integration of Block Operations for Data Invariant Automation of Digital Preprocessing and Analysis of Biological and Biomedical Raman Spectra

Schulze

Turner

2015

Appl Spectrosc

View full text Add to dashboard Cite

High-throughput information extraction from large numbers of Raman spectra is becoming an increasingly taxing problem due to the proliferation of new applications enabled using advances in instrumentation. Fortunately, in many of these applications, the entire process can be automated, yielding reproducibly good results with significant time and cost savings. Information extraction consists of two stages, preprocessing and analysis. We focus here on the preprocessing stage, which typically involves several steps, such as calibration, background subtraction, baseline flattening, artifact removal, smoothing, and so on, before the resulting spectra can be further analyzed. Because the results of some of these steps can affect the performance of subsequent ones, attention must be given to the sequencing of steps, the compatibility of these sequences, and the propensity of each step to generate spectral distortions. We outline here important considerations to effect full automation of Raman spectral preprocessing: what is considered full automation; putative general principles to effect full automation; the proper sequencing of processing and analysis steps; conflicts and circularities arising from sequencing; and the need for, and approaches to, preprocessing quality control. These considerations are discussed and illustrated with biological and biomedical examples reflecting both successful and faulty preprocessing.

show abstract

“…These issues are not only pure classification accuracy and false positive/negative rates but also comprise facets like interpretability and class representation, model size, classification guarantee and other [3,45,44]. In the following we will graze some of these aspects without any claim of completeness.…”

Section: Attempts To Improve the Classifier Performancementioning

confidence: 99%

“…Thus, classifier models have to be designed to handle different classification criteria. Besides these objectives also other criteria might be of interest like classifier model complexity, the interpretability of the results or the suitability for real time applications [3].…”

Section: Introductionmentioning

confidence: 99%

Aspects in Classification Learning - Review of Recent Developments in Learning Vector Quantization

Kaden¹,

Lange²,

Nebel³

et al. 2014

Foundations of Computing and Decision Sciences

View full text Add to dashboard Cite

Abstract. Classification is one of the most frequent tasks in machine learning. However, the variety of classification tasks as well as classifier methods is huge. Thus the question is coming up: which classifier is suitable for a given problem or how can we utilize a certain classifier model for different tasks in classification learning. This paper focuses on learning vector quantization classifiers as one of the most intuitive prototype based classification models. Recent extensions and modifications of the basic learning vector quantization algorithm, which are proposed in the last years, are highlighted and also discussed in relation to particular classification task scenarios like imbalanced and/or incomplete data, prior data knowledge, classification guarantees or adaptive data metrics for optimal classification.

show abstract

Classification in high-dimensional spectral data: Accuracy vs. interpretability vs. model size

Cited by 34 publications

References 10 publications

Metric learning for sequences in relational LVQ

Metric learning for sequences in relational LVQ

Development and Integration of Block Operations for Data Invariant Automation of Digital Preprocessing and Analysis of Biological and Biomedical Raman Spectra

Aspects in Classification Learning - Review of Recent Developments in Learning Vector Quantization

Contact Info

Product

Resources

About