A diagnostic unified classification model for classifying multi-sized and multi-modal brain graphs using graph alignment

Psychiatric diseases are bringing heavy burdens for both individual health and social stability. The accurate and timely diagnosis of the diseases is essential for effective treatment and intervention. Thanks to the rapid development of brain imaging technology and machine learning algorithms, diagnostic classification of psychiatric diseases can be achieved based on brain images. However, due to divergences in scanning machines or parameters, the generalization capability of diagnostic classification models has always been an issue. We propose Meta-learning with Meta batch normalization and Distance Constraint (M2DC) for training diagnostic classification models. The framework can simulate the train-test domain shift situation and promote intra-class cohesion, as well as inter-class separation, which can lead to clearer classification margins and more generalizable models. To better encode dynamic brain graphs, we propose a concatenated spatiotemporal attention graph isomorphism network (CSTAGIN) as the backbone. The network is trained for the diagnostic classification of major depressive disorder (MDD) based on multi-site brain graphs. Extensive experiments on brain images from over 3261 subjects show that models trained by M2DC achieve the best performance on cross-site diagnostic classification tasks compared to various contemporary domain generalization methods and SOTA studies. The proposed M2DC is by far the first framework for multi-source closed-set domain generalizable training of diagnostic classification models for MDD and the trained models can be applied to reliable auxiliary diagnosis on novel data.

Collaborative filtering aims at predicting a user's interest for a given item based on a collection of user profiles. This article views collaborative filtering as a problem highly related to information retrieval, drawing an analogy between the concepts of users and items in recommender systems and queries and documents in text retrieval.We present a probabilistic user-to-item relevance framework that introduces the concept of relevance into the related problem of collaborative filtering. Three different models are derived, namely, auser-based, anitem-based, and aunified relevance model, and we estimate their rating predictions from three sources: the user's own ratings for different items, other users' ratings for the same item, and ratings from different but similar users for other but similar items.To reduce the data sparsity encountered when estimating the probability density function of the relevance variable, we apply the nonparametric (data-driven) density estimation technique known as theParzen-window method(or kernel-based density estimation). Using a Gaussian window function, the similarity between users and/or items would, however, be based on Euclidean distance. Because the collaborative filtering literature has reported improved prediction accuracy when using cosine similarity, we generalize the Parzen-window method by introducing aprojection kernel.Existing user-based and item-based approaches correspond to two simplified instantiations of our framework. User-based and item-based collaborative filterings represent only a partial view of the prediction problem, where the unified relevance model brings these partial views together under the same umbrella. Experimental results complement the theoretical insights with improved recommendation accuracy. The unified model is more robust to data sparsity because the different types of ratings are used in concert.

In this article I use a theory of individual utility maximization to derive a unified model of electoral behavior that includes both candidate choices and turnout decisions. Compared to this new unified model, existing specifications for jointly considering turnout and vote choice are found to be theoretically or empirically lacking. I provide methods for testing my model in elections with two or three candidates, and I show that the parameters of these models can be estimated without difficulty using maximum likelihood techniques. Application of these unified models to the 1988 and 1992 American presidential elections illustrates the potential contrasts between unified models and models that consider only candidate choices.

This commentary addresses the modeling and final analytical path taken, as well as the terminology used, in the paper "Hierarchical diagnostic classification models: a family of models for estimating and testing attribute hierarchies" by Templin and Bradshaw (Psychometrika, doi: 10.1007/s11336-013-9362-0, 2013). It raises several issues concerning use of cognitive diagnostic models that either assume attribute hierarchies or assume a certain form of attribute interactions. The issues raised are illustrated with examples, and references are provided for further examination.

The U.S. Navy is considering medium voltage DC (MVDC) power systems as opposed to traditional AC power systems in order to accommodate modern shipboard systems: high power sensors, electronic warfare, and weapons systems. A digital twin of an MVDC naval power system is useful so that its operation can be better understood. In this work, a scaled-down demonstration model of a modern MVDC naval power system is studied through a series of experiments showing the operational benefits of the system when implemented with digital twin technologies. Two portable hybrid unified models of the scaled demonstration system were developed with incorporated machine learning techniques to show improved autonomous control. Because the unified models are portable, they may be developed in one tool, but then used with another tool on another platform. In a field setting, for example, a unified model may be simulated alongside an actual Navy ship to better understand its operation and vice versa.

Change occurs throughout the software lifecycle. Software Configuration Management tools and techniques provide the foundation to effectively control change. With a growing number of approaches combining models from different domains into one unified, integrated model, ([15], [12]), there is also an emerging demand for SCM techniques and methods that are able to support these unified models. Traditional SCM systems operating on the abstraction of a filesystem and managing change at the granularity of textual lines are not adequate for these requirements. We propose a novel approach to SCM for unified models combining product versioning, operation-based deltas and change packages. To demonstrate feasibility we have implemented our approach in Sysiphus a suite of tools for collaborating over Software Engineering artifacts represented in a unified model.

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) was employed to analyze differential serum and urine peptides in patients with small cell lung cancer (SCLC) and healthy individuals, and SCLC diagnostic classification models were constructed. Serum and urine samples from 72 patients with SCLC, age- and gender-matched with 72 healthy individuals, were divided into training and testing sets in a 3:1 ratio. Serum and urine peptides were extracted using copper ion-chelating nanomagnetic beads, and mass spectra were obtained using MALDI-TOF-MS. Peptide spectra for the training set were analyzed, and the classification model was constructed using ClinProTools (CPT). The testing set was used for blinded model validation. For training-set sera, 122 differential peptide signal peaks with a mass of 0.8–10 kDa were observed, and 19 peptides showed significantly different expression [P<0.0005; area under curve (AUC) ≥0.80]. CPT screened 5 peptide peaks (0.8114, 0.83425, 1.86655, 4.11133 and 5.81192 kDa) to construct the classification model. The testing set was used for the blinded validation, which had 95.0% sensitivity and 90.0% specificity. For the training-set urine, 132 differential peptide signal peaks with m/z ratios of 0.8–10 kDa were observed, and 8 peptides had significantly different expression (P<0.0005; AUC ≥0.80). Then, 5 peaks (1.0724, 2.37692, 2.7554, 4.75475 and 4.7949 kDa) were used for classification model construction. The testing set was used for 36 blinded validation, which had 85.0% sensitivity and 80.0% specificity. Among the differential peptides, 3 had the same significant peaks at 2.3764, 0.8778 and 0.8616 kDa, identified as fibrinogen α, glucose-6-phosphate isomerase and cyclin-dependent kinase-1, respectively. The present study highlighted the differences that exist in serum and urine peptides between patients with SCLC and healthy individuals. Serum and urine peptide diagnostic classification models could be constructed using MALDI-TOF-MS, and showed high sensitivity and specificity.

The management and disposal of waste is a severe social issue and an essential part of ecological sustainability. As an important component of the green, low-carbon, and recycling economic system, the identification and classification of recyclable waste is the premise of its reuse and energy conservation. The main issues at hand are to improve the classification accuracy and reliability of recyclable waste and to achieve automatic classification. The methods based on physical characteristics and image-based methods are inaccurate and unreliable. The current spectroscopy methods need to process the detected samples in advance, unsuitable for automatic detection. Based on material composition properties, the Laser-Induced Breakdown Spectroscopy (LIBS) technology is here proposed to accurately and reliably identify and classify recyclable waste into six categories at the level of consumer, such as paper, plastic, glass, metal, textile, and wood. The method is also used to subclassify the same category of waste for reuse at the level of a recycling factory. We subclassified metals into iron, stainless steel, copper, and aluminum and plastics into polyvinylchloride, polyoxymethylene, acrylonitrile-butadiene-styrene, polyamide, polyethylene, and polytetrafluoroethylene. The drop-dimension methods of LIBS spectra of waste were researched to eliminate noise and redundant information by principal component analysis (PCA) and linear discriminant analysis (LDA), respectively. Their clustering effects were analyzed to choose a suitable dimension. Combining the random forest (RF), back propagation neural network (BPNN), and convolutional neural network (CNN), we established and compared five classification models, PCA + RF, PCA + BPNN, LDA + RF, LDA + BPNN, and 1D-CNN. For the classification of six categories, the accuracies of proposed classification models are all more than 96%, and LDA(5D) + RF has 100% accuracy and optimal classification performance indices. For the subclassification of metals and plastics, PCA(8D) + RF has the highest classification accuracy of 98.77% and 99.52%, respectively.

Breast cancer remains one of the deadliest cancers for women, and timely detection is an important part in reducing mortality. Traditional mammography image analysis is met with low contrast, and a large fusion of breast tissues, especially in dens cases. To address the above issues, in this work deep learning-based approach that integrates CLAHE for local contrast enhancement, EfficientNet-B0 network for efficient and robust features extraction and rapid aggregation, as well as U-Net decoder to achieve accurate tumor segmentation is used. The model is trained and evaluated on the well-known public dataset Mini-MIAS, which comprises an archive of 322 mammogram images of various types of findings that are categorized as normal, benign findings or malignant. Our framework yielded an overall accuracy of 97.8% with precision, recall and $\mathbf{F 1}$ score values of 0.981,0.978 and 0.979, respectively that is higher as compared to that provided by conventional CNN and learning methods were achieved. By providing a unified CE, ROI segmentation, and classification model, the proposed framework is both reliable and interpretable, and stands to be practical for early tumor detection in the clinic.

In order to calculate the dynamic distribution of rail potential and stray current accurately, a multi‐state unified calculation model of rail potential and stray current is proposed, and the characteristics of different protection devices and reflux system in multi‐state DC railway systems are analysed. In the proposed model, the multi‐state characteristics of the traction substation, train, protection devices and reflux system are uniformly simulated by unified node models. The mixed simulation parameters of the catenary and reflux system are unified to lumped parameter networks, and the multi‐node voltage equations are established based on the simulation model. Then, the power flow of DC railway systems is calculated using an iterative solution, and the dynamic distribution of rail potential and stray current are solved by differential equations. The multi‐state unified calculation model is built and the distribution of rail potential and stray current in different conditions are simulated based on the parameters of Wuxi Subway Line 2. In order to verify the multi‐state unified calculation model proposed, field tests and simulations are carried out. Results show that the multi‐state unified calculation model can unify the complex states of DC railway systems and calculate the distribution of rail potential and stray current accurately.

Harvested plums can be stored at low temperatures for a long time. During storage, several quality attributes of plums undergo changes, which are also associated with their best-before date. This research proposed a fast method for the quick determination of plum storage time. With the integration of near-infrared spectra measurements, plum quality parameters (firmness, soluble solids content, pH, and color), several classification models have been developed to determine the storage time. A variety of methods such as linear discriminant analysis (LDA), partial least squares (PLS), support vector machine (SVM), and generalized linear models (GLM) were used. It is concluded that the models based on near-infrared spectra only with 14 selected features using LDA and PLS provide the best estimations, with accuracy values of 0.9917 and 0.9854, respectively. The models built with the attributes provide favorable results when using LDA and SVM (Radial basis), with accuracy values of 0.9261 and 0.8734, respectively. Novelty impact statement For the first time, we proposed the use of non-destructive NIRS to establish classification models in plums based on post-harvest storage time for maturity prediction. At the same time, several fruit quality attributes were also used for comparison.

This paper introduced a novel olfactory visualization freshness sensor array prepared by dropping pH indicators on UV lithography hydrophilic–hydrophobic paper under room temperature. The sensor array was applied in Grass carp spoilage classification. The sensor array can generate different discoloration patterns that respond to the pH change of the fish spoil; therefore, it can distinguish Grass carp freshness using pattern recognition. Classification models were built with linear discriminant analysis (LDA) and backpropagation artificial neural network (BP-ANN) after extracting the sensor array’s feature color information with principal component analysis (PCA). The performance of the BP-ANN model was superior to that of the LDA model, with an optimum recognition rate of 95.45% (training set) and 90.90% (testing set). In conclusion, with a suitable algorithm choice and rational pH indicator arrangement, this olfactory visualization freshness sensor array can potentially monitor food freshness. We prepared Grass carp freshness sensor array through dropping pH-indicators on hydrophilic–hydrophobic filter paper that fabricated by UV lithography under room temperature. The hydrophobic edges of the filter paper can maintain the sensor spots' shape and provide stable and consistent colorimetric change for Grass carp freshness recognition. Colorimetric classification models are built using LDA and BP-ANN based on pattern recognition, and the recognition rate of the BP-ANN model is above 90%. Novelty impact statement We prepared Grass carp freshness sensor array through dropping pH-indicators on hydrophilic-hydrophobic filter paper that fabricated by UV lithography under room temperature. The hydrophobic edges of the filter paper can maintain the sensor spots’ shape and provide stable and consistent colorimetric change for Grass carp freshness recognition. Colorimetric classification models are built using LDA and BP-ANN based on pattern recognition, and the recognition rate of the BP-ANN model is above 90 %.

The present study aimed at investigating invariance of a diagnostic classification model (DCM) for reading comprehension across gender. In contrast to models with continuous traits, diagnostic classification models inform mastery of a finite set of latent attributes, e.g., vocabulary or syntax in the reading context, and allow to provide fine grained feedback to learners and instructors. The generalized deterministic, noisy “and” gate (G-DINA) model was fit to item responses of 1000 male and female individuals to a high-stakes reading comprehension test. Use of the G-DINA model allowed for minimal assumption on the relationship of latent attribute profiles and item-specific response probabilities, i.e., the G-DINA model can represent compensatory or non-compensatory relationships of latent attributes and response probabilities. Item parameters were compared across the two samples, and only a small number of item parameters were statistically different between the two groups, corroborating the result of a formal measurement invariance test based on the multigroup G-DINA model. Neither correlations between latent attributes were significantly different across the two groups, nor mastery probabilities for any of the attributes. Model selection at item level showed that from among the 18 items that required multiple attributes, 16 items picked different rules (DCMs) across the groups. While this seems to suggest that the relationship among the attributes of reading comprehension differs across the two groups, a closer inspection of the rules picked by the items showed that almost in all cases the relationships were very similar. If a compensatory DCM was suggested by the G-DINA framework for an item in the female group, a model belonging to the same family resulted for the male group.

Learners and educators in the health professions have called for more fine-grained information (subscores) from assessments, beyond a single overall test score. However, due to concerns over reliability, there have been limited uses of subscores in practice. Recent advances in latent class analysis have made contributions in subscore reporting by using diagnostic classification models (DCMs), which allow reliable classification of examinees into fine-grained proficiency levels (subscore profiles). This study examines the innovative and practical application of DCM framework to health professions educational assessments using retrospective large-scale assessment data from the basic and clinical sciences: National Board of Medical Examiners Subject Examinations in pathology (n = 2,006) and medicine (n = 2,351). DCMs were fit and analyzed to generate subscores and subscore profiles of examinees. Model fit indices, classification (reliability), and parameter estimates indicated that DCMs had good psychometric properties including consistent classification of examinees into subscore profiles. Results showed a range of useful information including varying levels of subscore distributions. The DCM framework can be a promising approach to report subscores in health professions education. Consistency of classification was high, demonstrating reliable results at fine-grained subscore levels, allowing for targeted and specific feedback to learners.

Large‐scale dieback of ash trees (Fraxinus spp.) caused by the fungus Hymenoscyphus fraxineus is posing an immense threat to forest health in Europe, requiring effective monitoring at large scales. In this study, a pipeline was created to find ash trees and classify dieback severity using high‐resolution hyperspectral imagery of individual tree crowns (ITCs). Hyperspectral data were collected in four forest sites near Cambridge, UK, where 422 ITCs were manually delineated and labelled using field‐measurements of species and dieback severity (for ash trees). Four algorithms, namely linear discriminant analysis (LDA), principal components analysis coupled with LDA (PCA‐LDA), partial least squares discriminant analysis (PLS‐DA) and random forest (RF), were used to build classification models for species and dieback severity classification. The effect of dark‐pixel filtering on classification accuracy was evaluated. The best performing models were then coupled with automatic ITC segmentation to map species and ash dieback distribution over 16.8 hectares of woodland. We calculated and partitioned the coefficient of variation (CV) of the reflected ash spectra to find variable wavebands associated with dieback. PLS‐DA and LDA were most accurate for classifying ITC species identifies (overall accuracy &gt;90%), whereas RF was most accurate for classifying ash dieback severity (overall accuracy 77%). Dark pixel filtering further increased the accuracy of species classification (+6%), but not disease classification. The reflectances of narrow blue (415 nm), red‐edge (680 nm) and NIR (760 nm) bands had high CV across disease classes and should be included if multispectral imagery were to be used to monitor ash dieback. The study demonstrates the possibility of using remote sensing to forward epidemiological research by monitoring forest pathogens in landscape scales, which would allow temperate forest managers to control pathogen outbreaks, assess associated impacts and restore affected forests much more effectively.