Latent Correlations Research Articles

Intricate Supply Chain Demand Forecasting Based on Graph Convolution Network

Globalization has contributed to the increasing complexity of supply chain structures. In this regard, precise demand forecasting for the intricate supply chain holds paramount importance in effective supply chain management. Traditional statistical models and deep learning methods often face challenges in efficiently discerning correlations within a myriad of interconnected demands. To tackle this issue, this paper proposes an intricate supply chain demand forecasting method based on graph convolution networks adept at handling non-Euclidean data. First, the companies within the supply chain are treated as nodes in the graph structure, and the relationships between them are treated as edges, with demand data serving as the features of these edges. Then, a graph convolutional network is constructed to aggregate node and edge information. Through a multi-layer network, the relationships among nodes, edges, and historical demand are established to facilitate the prediction of supply chain demands. In this process, the graph convolutional network incorporates supply chain connectivity information into demand time series analysis. This integration of surface-level topological features and deeper latent correlation attributes across the supply chain’s nodes refines the demand forecasting precision across the entire supply chain. The validation experiment in this paper is grounded in sales data of a singular product from multiple sales nodes of an electronics company. The results demonstrate that the proposed method surpasses four other traditional demand forecasting algorithms significantly in terms of accuracy, providing substantial evidence for the superior performance of graph networks in the analysis of intricate supply chain relationships.

Cognitive, affective and sociological predictors of school performance in mathematics

Mathematics is a key school subject for some of the most lucrative and economically important careers. Low mathematics performance in school is associated with low psychometric intelligence, family socioeconomic status (SES), specific mathematical abilities, and high mathematics anxiety. We used a sample of Hungarian schoolchildren (N = 102, mean age = 12.3 years) to directly compare the predictive power of general intelligence, specific mathematical abilities measured by the Pedagogical Examination of Dyscalculia (DPV), mathematics anxiety, and socioeconomic status for mathematics grades. Mathematics grades correlated with IQ, specific mathematic ability, mathematics anxiety, and a composite measure of family SES. The WISC-IV showcased a manifest correlation of 0.62 and a latent correlation of 0.78 with the DPV and high manifest (r=-0.53) and latent (r=-0.59) correlations with mathematics anxiety. IQ alone accounted for 52% of the variance in mathematics grades. IQ, specific mathematical ability, family SES and mathematics anxiety jointly accounted for 56% of the variance in grades, with a non-significant contribution of specific mathematical ability and family SES over IQ and a marginal contribution of mathematics anxiety. Our results show that psychometric intelligence is the most important predictor of mathematics grades, while family SES and specific mathematical abilities are only associated with grades to the extent they reflect psychometric intelligence. The results, however, confirmed a small role of mathematics anxiety over intelligence in predicting grades.

Validating scores from the short form of the Music Self-Perception Inventory (MUSPI-S) with seventh- to ninth-grade school students in Germany

The Music Self-Perception Inventory-Short (MUSPI-S) is a theoretically based instrument consisting of 28 items with subscales that assess music self-concept globally and in six specific areas. The present study sought to validate a German translation of the MUSPI-S using a sample of 444 secondary school students. In addition to participants’ scores from the German translation of the MUSPI-S, associated data included demographic background variables, self-reported school grades in various subjects, and interest in the school subject of music. Confirmatory factor analyses supported the hypothesized factor structure and measurement model for MUSPI-S scores. Latent factor correlations among MUSPI-S subscale scores and correlations of these scores with school grades and interest in the school subject of music aligned well with theoretical expectations and previous research. Findings also provided evidence for strict measurement invariance of scores in the German MUSPI-S across subgroups of gender, school year (i.e., grade level), and musical activity. MUSPI-S subscale factor-level means did not reliably differ among seventh-, eighth-, and ninth-grade students but favored females over males and musically active over non-active students on nearly all subscales. Overall, this evidence supports the use of the German translation of the MUSPI-S in assessing a wide range of important self-perceived musical abilities for research and practical assessment purposes.

Who flourishes in school? The interplay of academic self-concept and personality and its role for academic performance in middle adolescence.

Why are some students more successful than others? We combined motivational and personality predictors and jointly examined the relevance of subject-specific academic self-concepts and Big Five personality traits for academic performance. Based on data from two independent studies of German 9th graders (Study 1: N = 1,508, Mage = 14.98 years, 51% female, 38% immigrant background; Study 2: N = 19,783, Mage = 15.10 years, 50% female, 36% immigrant background), we, first, estimated latent bivariate correlations to investigate the nomological net between these socioemotional characteristics. Second, using latent moderated regression models, we examined the role of the main and interaction effects of both characteristics for academic performance levels and changes assessed by grades and test scores. Finally, we tested whether the relevance of socioemotional characteristics for academic performance differed across sociodemographic characteristics. Five findings stand out: First, we established widely consistent nomological nets between the academic self-concepts and Big Five traits, especially regarding the German self-concept. Second, the domain-specific self-concepts were consistent predictors of different academic performance measures in the respective subject. Third, beyond the established main effects of openness and conscientiousness, all Big Five traits contributed to performance in some way. Fourth, despite some inconsistencies, socioemotional characteristics formed only synergistic interaction effects. Fifth, students' sociodemographic background was likewise important illustrating main effects on performance and foremost synergistic interaction effects with socioemotional predictors. Our results highlight the complex interplay between motivation, personality, and sociodemographic variables in predicting academic achievement and underline the need to be mindful of this interactive nature. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

2-Nitro-4-Carboxyphenylhydrazine and 2,4-Dicarboxylphenylhydrazine as a Pair of Novel Reactive MALDI Matrices for Rapid and Accurate Profiling of N-Glycome in Dual Ion Modes.

N-glycosylation is closely linked to a wide range of biological functions in organisms. Owing to the constriction of awful crystals formed by conventional MALDI matrices and the extremely inferior ionization efficiency of N-glycans, the traditional direct detection of N-glycans by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) has been gradually replaced by postderivatization detection using reactive matrices. Nevertheless, the laborious identification of complex spectral peaks remains the major difficulty in N-glycan profiling. Hence, we logically designed and synthesized two novel reactive matrices, 2-nitro-4-carboxylphenylhydrazine (NCPH) and 2,4-dicarboxylphenylhydrazine (DCPH), and separately combined them with the acidic matrix 2,5-dihydroxybenzoic acid (DHB) to constitute two composite matrices with high on-target derivatization efficiency and significant promotion of N-glycan ionization for productive MALDI analysis in dual ion modes. Using both composite matrices, we can actualize MALDI-MS and MS2 mass calibration in dual ion modes by postderivatization detection and fragmentation of dextrans and selectively enhance the ionization effect of oligosaccharides in mixed systems. Quite homogeneous cocrystals can ensure N-glycan quantification with decent linearity and reproducibility. A fixed mass difference derived from the identical N-glycan in two ion modes is available for rapid identification in complex biological samples. Ultimately, the developed strategy was triumphantly employed to identify and quantify the relative content and alteration tendency of peach N-glycans, which can be referable to the latent correlation between N-glycan expression and peach ripening.

Atlas-Based Structural Disconnectomes Are Associated with Cognitive Performance in Brain Tumors.

Background: Brain tumors are associated with impaired cognitive functioning, which may result from disruptions in brain structural connectivity. Estimating structural disconnections is a more advantageous representation of tumor impact and can be performed indirectly through normative brain atlases. Materials and Methods: Using a publicly available dataset of glioma and meningioma patient MRI scans and tumor masks, latent correlations were estimated between measures of structural disconnection and attention-based cognitive functioning. These measures included gray matter (GM) parcel damage, white matter tract damage, GM parcel-to-parcel disconnections, and reaction time (RTI) as part of the Cambridge Neuropsychological Test Automated Battery to assess attention. Results: Preprocessing pipelines with two different methods of minimizing the pathology impact on MRI normalization were utilized: cost-function masking and lesion filling. The results across both pipelines were nearly consistent, with significant correlations mainly found between RTI measures and the damage to the left inferior fronto-occipital and uncinate fasciculus, as well as the left prefrontal-visual disconnections. Conclusions: This alludes to the importance of left-hemispheric prefrontal-visual coupling in attention-based tasks, particularly those involving object- and feature-based attention.

The apple does not fall far: stable predictive relationships between parents’ ratings of their own and their children’s self-regulatory abilities

Having control over your own behavior and impulses is a critical skill that influences children’s academic, social, and emotional development. This study investigates the stability and predictive relationships between parents’ ratings of their own and their children’s executive function and delay aversion. Using data from approximately 1700 families collected during the COVID-19 pandemic, we employed hierarchical structural equation models and cross-lagged panel models to analyze the temporal stability and directional influences of executive function and delay aversion assessments.Our analysis revealed a substantial latent correlation (r = 0.48, p < 0.001) between parents’ and children’s executive function problems, indicating a shared variance of approximately 23%. Significant cross-lagged effects were found, with parental executive functions at T1 predicting child executive functions at T2 (β = 0.16, p = 0.005). For delay aversion, we found a latent correlation of r = 0.53 (p < 0.001) and significant within-timepoint and temporal stability, but no cross-lagged effects.These findings suggest that higher levels of executive function problems reported by parents at T1 correspond to an increased perception of similar problems in their children at T2. This highlights the importance of parental self-perception in assessing children’s abilities. Our results highlight the importance of incorporating family dynamics into interventions targeting executive function difficulties and delay aversion in children, and understanding this interplay enables the development of more effective, individualized approaches to support positive developmental outcomes.

Maternal depressed mood and serotonergic antidepressant treatment during pregnancy differentially shape the continuity between fetal–newborn neurobehaviour

BackgroundPrenatal serotonin reuptake inhibitor (SRI) antidepressant exposure is associated with newborn neurobehavioural disturbances, but it remains unclear whether this reflects a transient pharmacologic condition or an altered neurodevelopmental trajectory emerging in utero from sustained gestational SRI exposure. AimThis study explored longitudinal relationships between third-trimester fetal physiology and newborn neurobehaviour, and determined whether early neurobehavioural continuity is shaped by prenatal SRI or depression exposure. MethodsParticipants were 127 pregnant mothers and their fetal-newborn offspring. Four groups were defined based on antenatal depressive symptoms and SRI treatment: Control (n = 51), Depressed (unmedicated; n = 35), SRI-Depressed (n = 26) and SRI-Non-Depressed (n = 15). Doppler measures of fetal heart rate (fHR), motor activity and vascular hemodynamics were obtained at 36-weeks' gestation, then newborn neurobehavioural maturity was evaluated at postnatal day-7. Partial least squares analysis was used to identify latent correlations between fetal-newborn measures; associations were further studied with hierarchical regression testing group moderation. ResultsTwo dimensions described 74 % of the covariance between fetal physiologic and newborn neurobehavioural measures (permuted p < 0.05). Three latent fetal-newborn relationships were significantly moderated by group: (1) lower fHR variability, and (2) greater fHR decelerations, predicted lower alertness/orientation scores but only in SRI-Depressed-group newborns; and (3) lower fetal cerebrovascular resistance predicted lower motor scores in Depressed-group newborns. SRI treatment to euthymia was not associated with fetal-newborn neurobehavioural disturbances. ConclusionsMaternal depression, both unmedicated and SRI-treated with persistent/poorly-managed mood symptoms, differentially shaped fetal-newborn neurobehavioural continuity. These findings suggest that neurobehavioural disturbances may predate birth, and underscore the importance of effective mental health management during pregnancy.

An unsupervised learning framework for health diagnosis by incorporating multiscale data

Integrating real-time evaluations of health states with Structural Health Monitoring (SHM) data and environmental observations is crucial, particularly under large datasets. Not only is the volume of real-time data analysis substantial, but traditional algorithms like wavelet analysis often require extensive prior work and significant manual input, limiting the application of SHM and big data technologies in structural analysis. To overcome this limitation, this study introduces an unsupervised analytical framework coupling a Denoising Sparse Wavelet Network (DeSpaWN) and a dynamic-inner Principal Component Analysis (DiPCA) algorithm. The DeSpaWN is designed to extract sparse representation from multiscale sensor signals, and the features are subsequently analyzed using the DiPCA to assess health diagnoses. The DeSpaWN is specifically tailored for monitoring high-frequency time series dynamically and adaptively, enabling reconstruction and extraction of signals across various frequency domains. Additionally, the DiPCA effectively handles large-scale time-series data, providing precise evaluations by accounting for dynamic latent correlations. To validate the method, a 5-Degree of Freedom (DoF) structure model is established, equipped with five virtual sensors at each DoF and an anemometer at the top DoF. The sensor signals, obtained under different health states and subjected to artificial excitations, are input into the proposed framework. Validation results demonstrate the effectiveness in synthesizing multi-sensor data and assessing structural integrity. The framework's practical utility is showcased through a real-world case study that assesses the structural integrity of a wind turbine. By integrating real and virtual sensor data collected over a year and cross-referencing this with the fault log, the study validates the algorithm's efficacy in processing multisource heterogeneous data and its precision in structural health diagnosis. The study substantiates real-time tracking of health states for SHM-equipped structures, underscoring its potential to enhance safety and reliability in civil infrastructure.

A short-term wind power prediction method via self-adaptive adjacency matrix and spatiotemporal graph neural networks

Graph neural networks (GNN) recently have been successfully applied in wind power prediction by employing geo-information to construct the graphs that serve as the GNN-based prediction model. However, these graphs cannot maintain the latent correlation and attribute of the wind turbines, leading to inferior performance. This research proposes an optimizing wind power prediction model through attention mechanism and spatiotemporal graph neural networks. Initially, the spectral clustering and a self-adjacency matrix to construct the graph nodes and edges. Subsequently, the proposed ASTGNN combined from graph convolutional networks and a gated recurrent unit with an attention mechanism is designed to act as the prediction model. A comprehensive set of experiments has been carried out and the proposed method not only improved prediction accuracy but also enhanced computational efficiency. According to the results obtained, the developed model demonstrated an improvement that is up to 13 % in terms of RMSE and 6.7 % in terms of computation time compared to the latest models.

Using the multidimensional nominal response model to model faking in questionnaire data: The importance of item desirability characteristics.

Faking in self-report personality questionnaires describes a deliberate response distortion aimed at presenting oneself in an overly favorable manner. Unless the influence of faking on item responses is taken into account, faking can harm multiple psychometric properties of a test. In the present article, we account for faking using an extension of the multidimensional nominal response model (MNRM), which is an item response theory (IRT) model that offers a flexible framework for modeling different kinds of response biases. Particularly, we investigated under which circumstances the MNRM can adequately adjust substantive trait scores and latent correlations for the influence of faking and examined the role of variation in the way item content is related to social desirability (i.e., item desirability characteristics) in facilitating the modeling of faking and counteracting its detrimental effects. Using a simulation, we found that the inclusion of a faking dimension in the model can overall improve the recovery of substantive trait person parameters and latent correlations between substantive traits, especially when the impact of faking in the data is high. Item desirability characteristics moderated the effect of modeling faking and were themselves associated with different levels of parameter recovery. In an empirical demonstration with N = 1070 test-takers, we also showed that the faking modeling approach in combination with different item desirability characteristics can prove successful in empirical questionnaire data. We end the article with a discussion of implications for psychological assessment.

A long-term dissolved oxygen prediction model in aquaculture using transformer with a dynamic adaptive mechanism

The prediction of Dissolved Oxygen (DO) in oceans is crucial for the survival of marine life and the management of marine farms. However, due to the inherent instability and uncertainty of DO, prediction models frequently necessitate parameter adjustments, posing challenges to maintaining reliable predictive performance and generalization capabilities. Based on this, a long-term dissolved oxygen prediction model in aquaculture using Transformer with a dynamic adaptive mechanism is proposed in the paper. A multi-scale latent correlation decomposition strategy was devised, merging both time and frequency domain decompositions. This approach proficiently untangles time series exhibiting intricate non-linear cycles, thereby delving deeper into and furnishing a more comprehensive array of feature data. Furthermore, an uncertainty-aware attention mechanism was posited, ensuring accurate pinpointing of pivotal data amidst a highly volatile data environment, thus cementing prediction accuracy. Finally, the incorporation of an adaptive hyperparameter adjustment mechanism enables the model to recalibrate in line with feature insights, bolstering its generalization potential across marine farms in diverse regions. Empirical results show that the model proposed in this study can effectively accomplish high-precision long-term forecasting. When juxtaposed with other models under comparable conditions, its mean squared error significantly diminished by 40.2%, the mean absolute error contracted by 55.4%, and concurrently, the R2 value ascended by 7%. This research proffers a fresh vantage point for long-term marine chemical data predictions and extends robust technical bolstering for refining and augmenting aquacultural strategies.

Latent level correlation modeling of multivariate discrete-valued financial time series

Latent level correlation modeling of multivariate discrete-valued financial time series

F-norm based low-power motion recognition on wearable devices in the presence of outlier motions

Motion recognition refers to the intelligent recognition of human motion using data collected from wearable sensors, which exceedingly has gained significant interest from both academic and industrial fields. However, temporary-sudden activities caused by accidental behavior pose a major challenge to motion recognition and have been largely overlooked in existing works. To address this problem, the multi-dimensional time series of motion data is modeled as a Time-Frequency (TF) tensor, and the original challenge is transformed into a problem of outlier-corrupted tensor pattern recognition, where transient sudden activity data are considered as outliers. Since the TF tensor can capture the latent spatio-temporal correlations of the motion data, the tensor MPCA is used to derive the principle spatio-temporal pattern of the motion data. However, traditional MPCA uses the squared F-norm as the projection distance measure, which makes it sensitive to the presence of outlier motion data. Therefore, in the proposed outlier-robust MPCA scheme, the F-norm with the desirable geometric properties is used as the distance measure to simultaneously mitigate the interference of outlier motion data while preserving rotational invariance. Moreover, to reduce the complexity of outlier-robust motion recognition, we impose the proposed outlier-robust MPCA scheme on the traditional MPCANet which is a low-complexity deep learning network. The experimental results show that our proposed outlier-robust MPCANet can simultaneously improve motion recognition performance and reduce the complexity, especially in practical scenarios where the real-time data is corrupted by temporary-sudden activities.

Joint diffusion: mutual consistency-driven diffusion model for PET-MRI co-reconstruction

Objective. Positron Emission Tomography and Magnetic Resonance Imaging (PET-MRI) systems can obtain functional and anatomical scans. But PET suffers from a low signal-to-noise ratio, while MRI are time-consuming. To address time-consuming, an effective strategy involves reducing k-space data collection, albeit at the cost of lowering image quality. This study aims to leverage the inherent complementarity within PET-MRI data to enhance the image quality of PET-MRI. Approach. A novel PET-MRI joint reconstruction model, termed MC-Diffusion, is proposed in the Bayesian framework. The joint reconstruction problem is transformed into a joint regularization problem, where data fidelity terms of PET and MRI are expressed independently. The regular term, the derivative of the logarithm of the joint probability distribution of PET and MRI, employs a joint score-based diffusion model for learning. The diffusion model involves the forward diffusion process and the reverse diffusion process. The forward diffusion process adds noise to transform a complex joint data distribution into a known joint prior distribution for PET and MRI simultaneously, resembling a denoiser. The reverse diffusion process removes noise using a denoiser to revert the joint prior distribution to the original joint data distribution, effectively utilizing joint probability distribution to describe the correlations of PET and MRI for improved quality of joint reconstruction. Main results. Qualitative and quantitative improvements are observed with the MC-Diffusion model. Comparative analysis against LPLS and Joint ISAT-net on the ADNI dataset demonstrates superior performance by exploiting complementary information between PET and MRI. The MC-Diffusion model effectively enhances the quality of PET and MRI images. Significance. This study employs the MC-Diffusion model to enhance the quality of PET-MRI images by integrating the fundamental principles of PET and MRI modalities and leveraging their inherent complementarity. Furthermore, utilizing the diffusion model to learn the joint probability distribution of PET and MRI, thereby elucidating their latent correlation, facilitates a more profound comprehension of the priors obtained through deep learning, contrasting with black-box prior or artificially constructed structural similarities.

Two Effective Degrees of Freedom Can Represent the Dominant Features of Global Rayleigh Wave Dispersion Maps

AbstractObjectively exploring global variations in crust and upper mantle structure helps constrain fundamental aspects of Earth's plate tectonic and convective processes. Here we adopted a Variational Auto‐Encoder to explore the degrees of freedom of global Rayleigh wave dispersion maps at 4–40 mHz. We found that two latent variables sufficiently represent the global variations, suggesting inherent coupling between crustal and mantle seismic properties. We propose that the two extracted latent variables mostly correspond with crustal thickness and upper mantle thermal structure. The first variable shows low values for continental mountain belts and ocean spreading ridges, contrasted by high values for abyssal plains. The second variable shows low values for most oceanic lithosphere and Phanerozoic continental areas, contrasted by high values for Archean cratons. Latent space correlations indicate that continental lithosphere has more strongly coupled depth features than beneath the oceans, which might be a consequence of its longevity.

Aspects of Identity Questionnaire-IV: An Examination of Structural Validity, Gender Invariance, and Relationships with Mental Health and Basic Psychological Needs Among Adolescents

The present research evaluated evidence for structural and convergent validity and measurement invariance across gender of the Aspects of Identity Questionnaire-IV (AIQ-IV), which is designed to measure four identity orientations: Personal, Relational, Public, and Collective. We recruited two independent samples of Serbian adolescents aged 15–19 years (n 1 = 436, n 2 = 452). Both exploratory structural equation modeling (ESEM) and confirmatory factor analyses (CFA) were applied to examine the factor structure of the AIQ-IV. The ESEM model with four factors demonstrated a superior fit to the data in both samples and resulted in well-defined factors and deflated latent factor correlations compared to the CFA model. Multiple-group analysis supported the scalar invariance of the four-factor ESEM model across gender. The evidence for the convergent validity of the ESEM AIQ-IV factors was supported by meaningful correlations with various mental health indicators (life satisfaction, positive affect, negative affect, internalizing behaviors, and externalizing behaviors) and basic psychological needs (autonomy, relatedness, and competence) satisfaction and frustration. Our findings support the use of AIQ-IV as a multidimensional measure of identity orientations among adolescents and confirm the usefulness of applying the ESEM framework to evaluate the validity aspects of this scale.

Multimodal Image Classification by Multiview Latent Pattern Extraction, Selection, and Correlation.

The large amount of data available in the modern big data era opens new opportunities to expand our knowledge by integrating information from heterogeneous sources. Multiview learning has recently achieved tremendous success in deriving complementary information from multiple data modalities. This article proposes a framework called multiview latent space projection (MVLSP) to integrate features extracted from multiple sources in a discriminative way to facilitate binary and multiclass classifications. Our approach is associated with three innovations. First, most existing multiview learning algorithms promote pairwise consistency between two views and do not have a natural extension to applications with more than two views. MVLSP finds optimum mappings from a common latent space to match the feature space in each of the views. As the matching is performed on a view-by-view basis, the framework can be readily extended to multiview applications. Second, feature selection in the common latent space can be readily achieved by adding a class view, which matches the latent space representations of training samples with their corresponding labels. Then, high-order view correlations are extracted by considering feature-label correlations. Third, a technique is proposed to optimize the integration of different latent patterns based on their correlations. The experimental results on the prostate image dataset demonstrate the effectiveness of the proposed method.

The impact of bank money on stock market integration: evidence from the Eurozone

This paper explores the impact of monetary policy on the integration of the equity markets of the five largest economies of the Eurozone. We show that a rise in leverage and accommodating monetary policies significantly affects the probability of simultaneous stock market crashes. Moreover, we find that an increasing loan-to-base ratio divergence has a destabilizing effect on equity markets. Our findings are instrumental for international investors, who might operate under the assumption that Eurozone stock markets exhibit stable correlations during non-crisis periods, bolstered by the prevailing monetary regime. We suggest implementing a probability threshold as a macroprudential tool to manage latent correlations and ensure effective risk management.

Parallel Hierarchical Multi-View Feature Fusion Based on Canonical Correlation Analysis for Mammogram Retrieval

Due to the diversity of image sources, content-based multi-source image fusion and retrieval have shown promising capabilities in computer vision tasks, and especially when applied in Computer-Aided Diagnosis (CAD) to automate and improve the accuracy of medical image analysis. The combination of computer vision and CAD systems has the potential to revolutionize healthcare by augmenting the expertise of clinicians, improving overall diagnostic accuracy and helping experts in the clinical decision-making process by classifying and retrieving similar annotated clinical images to a given query. In the context of multi-view mammography interpretation, the concept of multi-view feature fusion has recently been studied to improve retrieval performance while effectively guaranteeing the complementarity of both MLO and CC views. However, conventional multi-view feature fusion makes descriptors long and lacks to take into consideration the relationship between descriptors. To deal with this issue, we propose two hierarchical multi-view feature fusion methods, for multi-view mammogram retrieval, based on the Canonical Correlation Analysis (CCA), which is the most commonly used multivariate parametric test. In fact, we have adapted CCA to determine the relationship between two descriptors by processing latent correlation factors. Moreover, after extracting descriptors for each view, a comparative study of texture and shape fusion descriptors is proposed in order to identify the more discriminative features for multi-view mammogram retrieval. Then, a query-dependent distance metric preserving both visual resemblance and semantic similarity is carried out to dynamically determine the more appropriate distance measure for each query image. Extensive experiments on the challenging Curated Breast Imaging Subset of Digital Database for Screening Mammography (CBIS-DDSM) have demonstrated the effectiveness of the proposed hierarchical multi-view feature fusion for mammogram retrieval, which outperforms the performance achieved either by conventional fused information or by single view information. To improve the transparency of our paper, the source code of the proposed method and the related dataset (including readme files) are publicly accessible through the following GitHub link: https://github.com/ABDERRAHIMMAR/Multi-View-Feature-Fusion-for-Mammogram-Retrieval . This open-access resource empowers researchers and practitioners to delve deeper into our methodology, fostering collaboration and advancements in the field of computer-aided diagnosis and medical image analysis.