A GA Based Measurement Method for Discontinuous Curve Fitting

Chen, Hung Jen; Chueh, Hao En; Huang, Deng

doi:10.4028/www.scientific.net/kem.460-461.106

Cited by 3 publications

(4 citation statements)

References 14 publications

(17 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They are presented in Table I, along with related information and references to methods with full automation potential. 39,–118…”

Section: Theorymentioning

confidence: 99%

“…95,126,129 However, peak parameter estimation requires initial estimates of the very same parameters; these can be difficult to gauge, 94 and unless guesses are very good, variations in initial parameters can lead to widely varying results. 93,130 The requisite initial estimates can be based on derivatives, 94 principal component analysis, 99 two-dimensional peak finding, 93 biologically inspired procedures, 92,100 monitoring changes in fitting errors as a function of minimum separable peak-peak distances, 95 and wavelet transforms, 97 among others. Both Lorentzian distributions (also known as Student's t-distributions with one degree of freedom or Cauchy distributions or Breit-Wigner distributions) and Gaussian distributions have been used for vibrational band curve fitting.…”

mentioning

confidence: 99%

See 1 more Smart Citation

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

“…They are presented in Table I, along with related information and references to methods with full automation potential. 39,–118…”

Section: Theorymentioning

confidence: 99%

mentioning

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

show abstract

“…Methods used to estimate the number of bands include derivative approaches 20 and principal component analysis. 22 Methods used to estimate initial band parameters include two-dimensional peak finding, 15 biologically inspired parameter-finding procedures, 23,24 monitoring changes in fitting errors as a function of minimum separable peak–peak distances, 19 Fourier self-deconvolution and maximum-likelihood restoration prior to curve fitting, 25 simultaneous deconvolution and curve-parameter extraction with adaptive Tikhonov regularizaron, 26 and wavelet transforms. 27 Regarding band shapes, both Lorentzian distributions (also known as Student's t -distributions with one degree of freedom, or Cauchy or Breit–Wigner distributions) and Gaussian distributions have been used for curve fitting.…”

Section: Introductionmentioning

confidence: 99%

Fully Automated Decomposition of Raman Spectra into Individual Pearson's Type VII Distributions Applied to Biological and Biomedical Samples

et al. 2015

View full text Add to dashboard Cite

Rapid technological advances have made the acquisition of large numbers of spectra not only feasible, but also routine. As a result, a significant research effort is focused on semi-automated and fully automated spectral processing techniques. However, the need to provide initial estimates of the number of peaks, their band shapes, and the initial parameters of these bands presents an obstacle to the full automation of peak fitting and its incorporation into fully automated spectral-preprocessing workflows. Moreover, the sensitivity of peak-fit routines to initial parameter settings and the resultant variations in solution quality further impede user-free operation. We have developed a technique to perform fully automated peak fitting on fully automated preconditioned spectra-specifically, baseline-corrected and smoothed spectra that are free of cosmic-ray-induced spikes. Briefly, the tallest peak in a spectrum is located and a Gaussian peak-fit is performed. The fitted peak is then subtracted from the spectrum, and the procedure is repeated until the entire spectrum has been processed. In second and third passes, all the peaks in the spectrum are fitted concurrently, but are fitted to a Pearson Type VII model using the parameters for the model established in the prior pass. The technique is applied to a synthetic spectrum with several peaks, some of which have substantial overlap, to test the ability of the method to recover the correct number of peaks, their true shape, and their appropriate parameters. Finally the method is tested on measured Raman spectra collected from human embryonic stem cells and samples of red blood cells.

show abstract

“…Least squares method (LSM) is a common method used to deal with optimization problems, but the obtained solutions are sometime local optimum solutions [2,9,13,14,15,18]. Previous researcher has demonstrated that genetic algorithm (GA) can effectively solve c optimization problems [3], but the coding method itself has an influence on the final results. In addition, the difficulty of parameter coding is also widely encountered in computational processes.…”

Section: Introductionmentioning

confidence: 99%

Applying Real-Valued Genetic Algorithm on Curve Fitting Problem

Chen¹,

Chueh

2010

AMR

View full text Add to dashboard Cite

Curve fitting refers to the process of finding an appropriate function that fits a finite set of data points. Representing a set of data points by a function is quite beneficial in data analysis and reapplication, and this technique is often used in engineering and technical problems. Fitting accuracy and computational time are usually the most crucial factors to be taken care of in curve fitting problems. Previous researchers have demonstrated that genetic algorithms can effectively solve curve fitting problems, but the difficulty of parameter coding is also widely encountered in computational processes. Hence, this study addresses on applying real-valued genetic algorithm to deal with curve fitting problems. Detailed discussion is made on the optimization efficiency among various data, and finally, some key parameters to curve fitting results are found and presented.

show abstract

A GA Based Measurement Method for Discontinuous Curve Fitting

Cited by 3 publications

References 14 publications

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

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

Fully Automated Decomposition of Raman Spectra into Individual Pearson's Type VII Distributions Applied to Biological and Biomedical Samples

Applying Real-Valued Genetic Algorithm on Curve Fitting Problem

Contact Info

Product

Resources

About