Self-Optimized One-Class Classification Using Sum of Ranking Differences Combined with a Receiver Operator Characteristic Curve

Abstract

A significant and common problem in analytical chemistry is determining if a sample belongs to a specific class, e.g., establishing if a food product is genuine or counterfeit or a tissue sample is benign or malignant. This problem is termed one-class classification (class modeling). Problematic with class modeling is determining which one-class classifier to use followed by the challenge of optimizing the chosen classifier (identifying the best tuning parameter value(s)). With spectroscopic data, two other conundrums arise: which data preprocessing method(s) and spectral region(s) to use. Presented in this paper is a hybrid fusion process that can combine nonoptimized classifiers across multiple instruments, preprocessing methods, and measurements. Instead of optimizing classifiers, a window of tuning parameters is used for each classifier. The flexible fusion method of sum of ranking differences (SRD) is applied to combine all assessment values. Defining the best SRD ranking value (threshold) for determining class membership is the one tuning parameter value needed. However, this SRD ranking value is automatically optimized by using a receiver operator characteristic (ROC) curve. The approach is demonstrated on two analytical data sets. The first is a beer authentication sample set measured on five instruments: near-infrared, mid infrared (MIR), ultraviolet, visible, and thermogravimetric. Three different fusion protocols of all five instruments are demonstrated. The second data set is MIR spectra of strawberry puree with two categories: strawberry puree and nonstrawberry puree. Fusing nonoptimized classifiers provides reliable classifications relative to accuracy, sensitivity, and specificity.

BHCNet: Neural Network-Based Brain Hemorrhage Classification Using Head CT Scan

“…There were 12,852 CT radiologist reports used for training and testing purpose. The performance is evaluated by using a receiver operator curve (ROC) [59], [60] and achieved 0.94 area under the curve (AUC).…”

Modeling with Multiple Correlated Spectral Data Based on Approximating the Nonlinear Spectrum Induced by Scattering

“…26 The fusion strategy is to directly integrate the preprocessing methods, rather than optimizing each preprocessing result. 27,28 Lemos and Kalivas 28 proposed a flexible fusion method to solve the selection problem of the processing methods and spectral regions, and to combine all assessment values with the sum of ranking difference (SRD). In order to suppress the effect of scattering, the combination of MSC and other pre-processing methods.…”

Interpretable One-Class Classification of Raman Spectra Using Prediction Bands Estimated by Wavelet Regression

“…4,6 The most prominent of such distancebased OCC methods is undoubtedly the soft independent modeling of class analogy (SIMCA). 5,7,8 Recently, ref 9 introduced a band-based OCC paradigm for spectral data, which uses, as a classification rule, the statistical concept of prediction band for future spectra in the wavelengths' space. This prediction band leverages the functional nature of the data, i.e., the fact that a spectrum is a discretized and noisy observation of a random smooth function of the wavelengths.…”

“…The recent technological advances in instrumentation, including miniaturized measurement devices and the possibilities of cloud-based computing systems, have made Raman spectroscopy practically accessible to many fields and applications. − Especially, in the quality by design context in the pharmaceutical industry, Raman spectroscopy is increasingly used to address qualitative analytical compliance testing problems, including the identification of raw materials and the verification of the compliance of the quality of intermediate drug products and bioprocesses with a reference quality. − In such problems, only representative spectra of the targeted identity or quality are available, while nontargeted or undesired identities or quality profiles are theoretically unlimited or cannot be representatively sampled. Hence, such qualitative testing problems are mathematically addressed by predicting the conformity of the Raman spectral features of unknown test samples to those of the representative reference set, based on classification rules defined using that reference set exclusively. , The type of mathematical techniques involved in such compliance verification tasks are known as “rigorous” one-class classification (OCC) methods. − Alternatively, when there also exist samples of some nontarget quality or products, even if these samples are not representative of all of the scenarios of nontarget quality or product profiles that might be encountered in the future, they can be used together with the reference set to define the classification rules. In the latter case, the resulting type of models are termed “compliant” OCC methods. − …”

“…Hence, such qualitative testing problems are mathematically addressed by predicting the conformity of the Raman spectral features of unknown test samples to those of the representative reference set, based on classification rules defined using that reference set exclusively. , The type of mathematical techniques involved in such compliance verification tasks are known as “rigorous” one-class classification (OCC) methods. − Alternatively, when there also exist samples of some nontarget quality or products, even if these samples are not representative of all of the scenarios of nontarget quality or product profiles that might be encountered in the future, they can be used together with the reference set to define the classification rules. In the latter case, the resulting type of models are termed “compliant” OCC methods. − …”