Search

These papers collectively showcase the adaptability and effectiveness of data mining techniques, making substantial contributions to the broader realm of " I Methods and Applications of Data Mining in Business Domains i ". In a business context, the challenge is that one would like to see (i) how the algorithms can be repeatable in the real world, (ii) how the patterns mined can be utilized by the business, and (iii) how the resulting model can be understood and utilized in the business environment [[1]]. Additionally, they provide insights into factors influencing the adoption of business intelligence systems (BISs) in small and medium-sized enterprises (SMEs) [[26]], and conduct a systematic literature review on AI-based methods for automating business processes and decision support [[27]]. This Special Issue invited researchers to contribute original research in the field of data mining, particularly in its application to diverse domains, like healthcare, software development, logistics, and human resources. [Extracted from the article]

Background: MicroRNAs (miRNAs) emerge in various organisms, ranging from viruses to humans, and play crucial regulatory roles within cells, participating in a variety of biological processes. In numerous prediction methods for miRNA-disease associations, the issue of over-dependence on both similarity measurement data and the association matrix still hasn't been improved. In this paper, a miRNA-Disease association prediction model (called TP-MDA) based on tree path global feature extraction and fully connected artificial neural network (FANN) with multi-head self-attention mechanism is proposed. The TP-MDA model utilizes an association tree structure to represent the data relationships, multi-head self-attention mechanism for extracting feature vectors, and fully connected artificial neural network with 5-fold cross-validation for model training. Results: The experimental results indicate that the TP-MDA model outperforms the other comparative models, AUC is 0.9714. In the case studies of miRNAs associated with colorectal cancer and lung cancer, among the top 15 miRNAs predicted by the model, 12 in colorectal cancer and 15 in lung cancer were validated respectively, the accuracy is as high as 0.9227. Conclusions: The model proposed in this paper can accurately predict the miRNA-disease association, and can serve as a valuable reference for data mining and association prediction in the fields of life sciences, biology, and disease genetics, among others. [ABSTRACT FROM AUTHOR]

Lower-dimensional representation of scientific text has attracted much attention among researchers due to its impact on many data mining and recommendation tasks. This paper studies two main research streams in scientific literature representation. First, both local and distributed representation viewpoints are reviewed and their advantages and disadvantages in lower dimensional representation are discussed. The paper then proposes a novel hybrid distributed technique for text representation. Using scientific articles as the major source of textual information, both the article's content and citation network are used to build a distributed and universal lower dimensional representation. The superiority of the new technique to the traditional methods is then justified in predicting the existence of links in large citation graphs. [ABSTRACT FROM AUTHOR]

Diabetic Data Analysis is a field of research which comes under analytics. Analytics is a subject of statistics to the extent that we read raw data by using computational techniques and then we make sense out of this raw data this is called analysis. An essential function in data mining and analytics is the Data Classification. A machine learning tool known as neural network is capable to perform various tasks in diabetic data analysis. Today, healthcare industries having large amount of data and to access that data analysis process is required, so there arise many complexities. Medicare industries face different kind of challenges, so it is very important to develop data analytics. In this paper an integrated approach is used to predict diabetes from neural network. Neural network can be taken as ubiquitous indicator. From various resources raw data has been collected and compare it to a tool that can be a trained machine for the prediction of diabetes patients. Main aim of integrating approach in neural network is to increase the accurate results in the prediction of diabetic patients. Big data is an approach to resolve the problem in an enhanced manner. A modeling structure is used in this paper. [ABSTRACT FROM AUTHOR]

Rapid mapping demands efficient methods for a fast extraction of information from satellite data while minimizing data requirements. This paper explores the potential of deep learning for the generation of high-resolution urban elevation data from Synthetic Aperture Radar (SAR) imagery. In order to mitigate occlusion effects caused by the side-looking nature of SAR remote sensing, two SAR images from opposing aspects are leveraged and processed in an end-to-end deep neural network. The presented approach is the first of its kind to implicitly handle the transition from the SAR-specific slant range geometry to a ground-based mapping geometry within the model architecture. Comparative experiments demonstrate the superiority of the dual-aspect fusion over single-image methods in terms of reconstruction quality and geolocation accuracy. Notably, the model exhibits robust performance across diverse acquisition modes and geometries, showcasing its generalizability and suitability for height mapping applications. The study's findings underscore the potential of deep learning-driven SAR techniques in generating high-quality urban surface models efficiently and economically. [ABSTRACT FROM AUTHOR]

Copyright of Inzynieria Mineralna is the property of Polskie Towarzystwo Przerobki Kopalin and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

We are pleased to present a special issue of Data Science and Engineering (DSE), which contains a collection of eight extended papers from the ICDM 2019 conference. The eight extended papers in this special issue cover a variety of topics related to data science and engineering. The seventh paper "Contextual Sentiment Neural Network for Document Sentiment Analysis" proposes a novel neural network model that can explain the process of its sentiment analysis prediction in a way that humans find natural and agreeable and can catch-up the summary of the contents. [Extracted from the article]

Entity-relationship extraction is a fine-grained task for constructing a knowledge graph of food public opinion in the field of food public opinion, and it is also an important research topic in the field of current information extraction. This paper aims at the multi-entity-to-relationship problem that often occurs in food public opinion, the entity-relationship types are extracted from the BERT (Bidirectional Encoder Representation from Transformers) network model; In the bidirectional long short-term memory network (BLSTM), the entity-relationship types extracted by BERT model are integrated, and the semantic role attention mechanism based on position awareness is introduced to construct a model BERT-BLSTM-based entity-relationship extraction model for food public opinion at the same time. In this paper, comparative experiments were conducted on the food sentiment data set. The experimental results show that the accuracy of the BERT-BLSTM-based food sentiment entity-relationship extraction model proposed in this paper is 8.7 ~ 13.94% higher than several commonly used deep neural network models on the food sentiment data set, which verifies the rationality and effectiveness of the model proposed in this paper. [ABSTRACT FROM AUTHOR]

The aim of this paper is to mine the information contained in the bridge health monitoring data as well as to improve the shortcomings of traditional identification methods. In this paper, a bridge damage identification method based on the combination of data mining and deep neural networks is introduced. Firstly, a noise reduction method based on parameter optimisation of wavelet threshold decomposition is proposed, which further removes the noise signal by introducing two adjustment parameters in the threshold function to adapt to different wavelet decomposition layers. Furthermore, the Fast Fourier Transform is used to analyse the feature pattern of the original signal in the frequency domain, and the modal frequency features that exhibit the difference in damage categories are extracted from the spectrogram through sliding windows. Finally, a large number of irrelevant variables with small weight contributions are discarded by principal component analysis, and only the sensitive features with the most informative categories are retained as the input to the deep neural networks. The experimental results show that the new metrics after the feature engineering process improve the ability of damage identification and have stronger robustness, while our damage identification scheme achieves a good balance between the model computation and recognition accuracy. Furthermore, the recognition accuracy of the deep neural networks reaches over 93% with only three feature dimensions retained. [ABSTRACT FROM AUTHOR]

Dark web vendor identification can be seen as an authorship aliasing problem, aiming to determine whether different accounts on different markets belong to the same real-world vendor, in order to locate cybercriminals involved in dark web market transactions. Existing open-source datasets for dark web marketplaces are outdated and cannot simulate real-world situations, while data labeling methods are difficult and suffer from issues such as inaccurate labeling and limited cross-market research. The problem of identifying vendors' multiple identities on the dark web involves a large number of categories and a limited number of samples, making it difficult to use traditional multiclass classification models. To address these issues, this paper proposes a metric learning-based method for dark web vendor identification, collecting product data from 21 currently active English dark web marketplaces and using a multi-dimensional feature extraction method based on product titles, descriptions, and images. Using pseudo-labeling technology combined with manual labeling improves data labeling accuracy compared to previous labeling methods. The proposed method uses a Siamese neural network with metric learning to learn the similarity between vendors and achieve the recognition of vendors' multiple identities. This method achieved better performance with an average F1-score of 0.889 and an accuracy rate of 97.535% on the constructed dataset. The contributions of this paper lie in the proposed method for collecting and labeling data for dark web marketplaces and overcoming the limitations of traditional multiclass classifiers to achieve effective recognition of vendors' multiple identities. [ABSTRACT FROM AUTHOR]

Purpose: The objective of this paper is to evaluate whether the data from consignors, logistics service providers (LSPs) and consignees contribute to the prediction of air transport shipment delays in a machine learning application. Design/methodology/approach: The research database contained 2,244 air freight intercontinental shipments to 4 automotive production plants in Latin America. Different algorithm classes were tested in the knowledge discovery in databases (KDD) process: support vector machine (SVM), random forest (RF), artificial neural networks (ANN) and k-nearest neighbors (KNN). Findings: Shipper, consignee and LSP data attribute selection achieved 86% accuracy through the RF algorithm in a cross-validation scenario after a combined class balancing procedure. Originality/value: These findings expand the current literature on machine learning applied to air freight delay management, which has mostly focused on weather, airport structure, flight schedule, ground delay and congestion as explanatory attributes. [ABSTRACT FROM AUTHOR]

Deep learning‐based models have become ubiquitous across a wide range of applications, including computer vision, natural language processing, and robotics. Despite their efficacy, one of the significant challenges associated with deep neural network (DNN) models is the potential risk of copyright leakage due to the inherent vulnerability of the entire model architecture and the communication burden of the large models during publishing. So far, it is still challenging for us to safeguard the intellectual property rights of these DNN models while reducing the communication time during model publishing. To this end, this paper introduces a novel approach using knowledge distillation techniques aimed at training a surrogate model to stand in for the original DNN model. To be specific, a knowledge distillation generative adversarial network (KDGAN) model is proposed to train a student model capable of achieving remarkable performance levels while simultaneously safeguarding the copyright integrity of the original large teacher model and improving communication efficiency during model publishing. Herein, comprehensive experiments are conducted to showcase the efficacy of model copyright protection, communication‐efficient model publishing, and the superiority of the proposed KDGAN model over other copyright protection mechanisms. [ABSTRACT FROM AUTHOR]

This paper proposes a novel packaging evaluation method using corrugated box images and an artificial neural network (ANN). An ANN works in a way similar to that of neurons in a human brain: by making connections between a trained dataset and the new images provided after training. The ANN has been implemented in the industry in various ways but limited in packaging evaluation. This paper is focused on the corrugated box damage prediction using ANN with the Orange Data Mining platform. By capturing the damaged corrugated box images with an ANN, damaged products can be identified allowing a decision to be made as to what type of package failure occurred. One of the benefits to using an ANN to evaluate corrugated box images is that it allows for the evaluation of package protection in a real distribution environment as compared to a controlled lab setting. In turn, this reduces the cost of testing, as the package failure will have been identified with the assistance of the ANN, rather than full retesting to identify where damage occurred. This process would also reduce costs associated with the usage of materials for testing, due to the lower number of test samples required. [ABSTRACT FROM AUTHOR]

Agricultural sustainability is dependent on the ability to predict crop yield, which is vital for farmers, consumers, and researchers. Most of the works used the amount of rainfall, average monthly temperature, relative humidity, etc. as inputs. In this paper, an attempt was made to predict the yield of the citrus crop (Washington Navel orange, Valencia orange, Murcott mandarin, Fremont mandarin, and Bearss Seedless lime) using weather factors and the accumulated heat units. These variables were used as input parameters in an artificial neural network (ANN) model. The necessary information was gathered during the growing seasons between 2010/2011 and 2021/2022 under Egyptian conditions. Weather factors were daily precipitation, yearly average air temperature, and yearly average of air relative humidity. A base air temperature of 13.0 °C was used to determine the accumulated heat units. The heat use efficiency (HUE) for cultivars was determined. The Bearss Seedless lime had the lowest HUE of 9.5 kg/ha °C day, while the Washington Navel orange had the highest HUE of 20.2 kg/ha °C day. The predictive performance of the ANN model with a structure of 9-20-1 with the backpropagation was evaluated using standard statistical measures. The actual and estimated yields from the ANN model were compared using a testing dataset, resulting in a value of RMSE, MAE, and MAPE of 2.80 t/ha, 2.58 t/ha, and 5.41%, respectively. The performance of the ANN model in the training phase was compared to multiple linear regression (MLR) models using values of R2; for MLR models for all cultivars, R2 ranged between 0.151 and 0.844, while the R2 value for the ANN was 0.87. Moreover, the ANN model gave the best performance criteria for evaluation of citrus yield prediction with a high R2, low root mean squared error, and low mean absolute error compared to the performance criteria of data mining algorithms such as K-nearest neighbor (KNN), KStar, and support vector regression. These encouraging outcomes show how the current ANN model can be used to estimate fruit yields, including citrus fruits and other types of fruit. The novelty of the proposed ANN model lies in the combination of weather parameters and accumulated heat units for accurate citrus yield prediction, specifically tailored for Egyptian regional citrus crops. Furthermore, especially in low- to middle-income countries such as Egypt, the findings of this study can greatly enhance the reliance on statistics when making decisions regarding agriculture and climate change. The citrus industry can benefit greatly from these discoveries, which can help with optimization, harvest planning, and postharvest logistics. We recommended furthering proving the robustness and generalization ability of the results in this study by adding more data points. [ABSTRACT FROM AUTHOR]

Today according to social media, the internet, Etc. Data is rapidly produced and occupies a large space in systems that have resulted in enormous data warehouses; the progress in information technology has significantly increased the speed and ease of data flow.text mining is one of the most important methods for extracting a useful model through extracting and adapting knowledge from data sets. However, many studies have been conducted based on the usage of deep learning for text processing and text mining issues.The idea and method of text mining are one of the fields that seek to extract useful information from unstructured textual data that is used very today. Deep learning and machine learning techniques in classification and text mining and their type are discussed in this paper as well. Neural networks of various kinds, namely, ANN, RNN, CNN, and LSTM, are the subject of study to select the best technique. In this study, we conducted a Systematic Literature Review to extract and associate the algorithms and features that have been used in this area. Based on our search criteria, we retrieved 130 relevant studies from electronic databases between 1997 and 2021; we have selected 43 studies for further analysis using inclusion and exclusion criteria in Section 3.2. According to this study, hybrid LSTM is the most widely used deep learning algorithm in these studies, and SVM in machine learning method high accuracy in result shown. [ABSTRACT FROM AUTHOR]

In the data retrieval process of the Data recommendation system, the matching prediction and similarity identification take place a major role in the ontology. In that, there are several methods to improve the retrieving process with improved accuracy and to reduce the searching time. Since, in the data recommendation system, this type of data searching becomes complex to search for the best matching for given query data and fails in the accuracy of the query recommendation process. To improve the performance of data validation, this paper proposed a novel model of data similarity estimation and clustering method to retrieve the relevant data with the best matching in the big data processing. In this paper advanced model of the Logarithmic Directionality Texture Pattern (LDTP) method with a Metaheuristic Pattern Searching (MPS) system was used to estimate the similarity between the query data in the entire database. The overall work was implemented for the application of the data recommendation process. These are all indexed and grouped as a cluster to form a paged format of database structure which can reduce the computation time while at the searching period. Also, with the help of a neural network, the relevancies of feature attributes in the database are predicted, and the matching index was sorted to provide the recommended data for given query data. This was achieved by using the Distributional Recurrent Neural Network (DRNN). This is an enhanced model of Neural Network technology to find the relevancy based on the correlation factor of the feature set. The training process of the DRNN classifier was carried out by estimating the correlation factor of the attributes of the dataset. These are formed as clusters and paged with proper indexing based on the MPS parameter of similarity metric. The overall performance of the proposed work can be evaluated by varying the size of the training database by 60%, 70%, and 80%. The parameters that are considered for performance analysis are Precision, Recall, F1-score and the accuracy of data retrieval, the query recommendation output, and comparison with other state-of-art methods. [ABSTRACT FROM AUTHOR]

In recent years, the use of deep neural network in effective network feature extraction and the design of efficient and high-precision hyperspectral image classification algorithms has gradually become a research hotspot for scholars. However, due to the difficulty of obtaining hyperspectral images and the high cost of annotation, the training samples are very limited. In order to cope with the small sample problem, researchers often deepen the network model and use the attention mechanism to extract features; however, as the network model continues to deepen, the gradient disappears, the feature extraction ability is insufficient, and the computational cost is high. Therefore, how to make full use of the spectral and spatial information in limited samples has gradually become a difficult problem. In order to cope with such problems, this paper proposes two-branch multiscale spatial–spectral feature aggregation with a self-attention mechanism for a hyperspectral image classification model (FHDANet); the model constructs a dense two-branch pyramid structure, which can achieve the high efficiency extraction of joint spatial–spectral feature information and spectral feature information, reduce feature loss to a large extent, and strengthen the model's ability to extract contextual information. A channel–space attention module, ECBAM, is proposed, which greatly improves the extraction ability of the model for salient features, and a spatial information extraction module based on the deep feature fusion strategy HLDFF is proposed, which fully strengthens feature reusability and mitigates the feature loss problem brought about by the deepening of the model. Compared with five hyperspectral image classification algorithms, SVM, SSRN, A2S2K-ResNet, HyBridSN, SSDGL, RSSGL and LANet, this method significantly improves the classification performance on four representative datasets. Experiments have demonstrated that FHDANet can better extract and utilise the spatial and spectral information in hyperspectral images with excellent classification performance under small sample conditions. [ABSTRACT FROM AUTHOR]

Currently, many fault diagnosis methods for rolling bearings based on deep learning are facing two main challenges. Firstly, the deep learning model exhibits poor diagnostic performance and limited generalization ability in the presence of noise signals and varying loads. Secondly, there is incomplete utilization of fault information and inadequate extraction of fault features, leading to the low diagnostic accuracy of the model. To address these problems, this paper proposes an improved dual-branch convolutional capsule neural network for rolling bearing fault diagnosis. This method converts the collected bearing vibration signals into grayscale images to construct a grayscale image dataset. By fully considering the types of bearing faults and damage diameters, the data are labeled using a dual-label format. A multi-scale convolution module is introduced to extract features from the data and maximize feature information extraction. Additionally, a coordinate attention mechanism is incorporated into this module to better extract useful channel features and enhance feature extraction capability. Based on adaptive fusion between fault type (damage diameter) features and labels, a dual-branch convolutional capsule neural network model for rolling bearing fault diagnosis is established. The model was experimentally validated using both Case Western Reserve University's bearing dataset and self-made datasets. The experimental results demonstrate that the fault type branch of the model achieves an accuracy rate of 99.88%, while the damage diameter branch attains an accuracy rate of 99.72%. Both branches exhibit excellent classification performance and display robustness against noise interference and variable working conditions. In comparison with other algorithm models cited in the reference literature, the diagnostic capability of the model proposed in this study surpasses them. Furthermore, the generalization ability of the model is validated using a self-constructed laboratory dataset, yielding an average accuracy rate of 94.25% for both branches. [ABSTRACT FROM AUTHOR]

Purpose: This paper aims to apply several data mining techniques for predicting the daily precision improvement of Jakarta Islamic Index (JKII) prices based on big data of symmetric volatility in Indonesia's Islamic stock market. Design/methodology/approach: This research uses big data mining techniques to predict daily precision improvement of JKII prices by applying the AdaBoost, K-nearest neighbor, random forest and artificial neural networks. This research uses big data with symmetric volatility as inputs in the predicting model, whereas the closing prices of JKII were used as the target outputs of daily precision improvement. For choosing the optimal prediction performance according to the criteria of the lowest prediction errors, this research uses four metrics of mean absolute error, mean squared error, root mean squared error and R-squared. Findings: The experimental results determine that the optimal technique for predicting the daily precision improvement of the JKII prices in Indonesia's Islamic stock market is the AdaBoost technique, which generates the optimal predicting performance with the lowest prediction errors, and provides the optimum knowledge from the big data of symmetric volatility in Indonesia's Islamic stock market. In addition, the random forest technique is also considered another robust technique in predicting the daily precision improvement of the JKII prices as it delivers closer values to the optimal performance of the AdaBoost technique. Practical implications: This research is filling the literature gap of the absence of using big data mining techniques in the prediction process of Islamic stock markets by delivering new operational techniques for predicting the daily stock precision improvement. Also, it helps investors to manage the optimal portfolios and to decrease the risk of trading in global Islamic stock markets based on using big data mining of symmetric volatility. Originality/value: This research is a pioneer in using big data mining of symmetric volatility in the prediction of an Islamic stock market index. [ABSTRACT FROM AUTHOR]

Financial fraud, considered as deceptive tactics for gaining financial benefits, has recently become a widespread menace in companies and organizations. Conventional techniques such as manual verifications and inspections are imprecise, costly, and time consuming for identifying such fraudulent activities. With the advent of artificial intelligence, machine-learning-based approaches can be used intelligently to detect fraudulent transactions by analyzing a large number of financial data. Therefore, this paper attempts to present a systematic literature review (SLR) that systematically reviews and synthesizes the existing literature on machine learning (ML)-based fraud detection. Particularly, the review employed the Kitchenham approach, which uses well-defined protocols to extract and synthesize the relevant articles; it then report the obtained results. Based on the specified search strategies from popular electronic database libraries, several studies have been gathered. After inclusion/exclusion criteria, 93 articles were chosen, synthesized, and analyzed. The review summarizes popular ML techniques used for fraud detection, the most popular fraud type, and evaluation metrics. The reviewed articles showed that support vector machine (SVM) and artificial neural network (ANN) are popular ML algorithms used for fraud detection, and credit card fraud is the most popular fraud type addressed using ML techniques. The paper finally presents main issues, gaps, and limitations in financial fraud detection areas and suggests possible areas for future research. [ABSTRACT FROM AUTHOR]

Cross-domain models based on Deep Neural Networks (Convolutional Neural Networks, Transformer Encoders and Attentional BLSTM models) have been tested, and also commercial systems for the same task (MeaningCloud, Microsoft Text Analytics and Google Cloud). The 15th International Work-Conference on Artificial Neural Networks (IWANN) is a biennial meeting seeks to provide a discussion forum for scientists, engineers, educators and students about the latest discoveries and realizations in the foundations, theory, models and applications of systems inspired on nature, using computational intelligence methodologies, as well as in emerging areas related to the above items. This typically involves various signal processing and machine learning steps used to transform raw sensor data into activity labels. [Extracted from the article]

Deep neural networks (DNNs), including convolutional neural networks (CNNs) and residual networks (ResNets) models, are able to learn abstract representations from the input data by considering a deep hierarchy of layers that perform advanced feature extraction. The combination of these models with visual attention techniques can assist with the identification of the most representative parts of the data from a visual standpoint, obtained through more detailed filtering of the features extracted by the operational layers of the network. This is of significant interest for analyzing remotely sensed hyperspectral images (HSIs), characterized by their very high spectral dimensionality. However, few efforts have been conducted in the literature in order to adapt visual attention methods to remotely sensed HSI data analysis. In this paper, we introduce a new visual attention-driven technique for the HSI classification. Specifically, we incorporate attention mechanisms to a ResNet in order to better characterize the spectral–spatial information contained in the data. Our newly proposed method calculates a mask that is applied to the features obtained by the network in order to identify the most desirable ones for classification purposes. Our experiments, conducted using four widely used HSI data sets, reveal that the proposed deep attention model provides competitive advantages in terms of classification accuracy when compared to other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

In order to solve the technical problem of fault signal recognition in the field of communication, this paper proposes an electronic communication fault signal recognition method based on data mining algorithm. Firstly, the K-means clustering algorithm is used to determine the cluster number k according to some attributes or class characteristics of the communication class samples, and the communication sample types are classified into a certain class so that the communication sample data in the cluster can be closely distributed and the data within a certain class range can be calculated by Euclidean distance formula. Then, this paper clusters the data. In the clustering data, BP neural network model is used to train and calculate the obtained clustering data again, which can map and deal with the complex nonlinear relationship between the fault information data of different clustering categories. The results show that the final error accuracy can be raised to about 20% by using the method in this paper. Conclusion. The algorithm designed in this paper can quickly predict the factors affecting the communication and find the communication fault information. [ABSTRACT FROM AUTHOR]

Purpose: The purpose of this paper is to apply various data mining techniques for predicting the financial performance of Islamic banking in Indonesia through the main exogenous determinants of profitability by choosing the best data mining technique based on the criteria of the highest accuracy score of testing and training. Design/methodology/approach: This paper used data mining techniques to predict the financial performance of Islamic banking by applying all of LASSO regression, random forest (RF), artificial neural networks and k-nearest neighbor (KNN) over monthly data sets of all the full-fledged Islamic banks working in Indonesia from January 2011 until March 2020. This study used return on assets as a real measurement of financial performance, whereas the capital adequacy ratio, asset quality and liquidity management were used as exogenous determinants of financial performance. Findings: The experimental results showed that the optimal task for predicting the financial performance of Islamic banking in Indonesia is the KNN technique, which affords the best-predicting accuracy, and gives the optimal knowledge from the financial performance of Islamic banking determinants in Indonesia. As well, the RF provides closer values to the optimal accuracy of the KNN, which makes it another robust technique in predicting the financial performance of Islamic banking. Research limitations/implications: This paper restricted modeling the financial performance of Islamic banking to profitability through the main determinants of return of assets in Indonesia. Future research could consider enlarging the modeling of financial performance using other models such as CAMELS and Z-Score to predict the financial performance of Islamic banking under data mining techniques. Practical implications: Owing to the lack of using data mining techniques in the Islamic banking sector, this paper would fill the literature gap by providing new effective techniques for predicting financial performance in the Islamic banking sector using data mining approaches, which can be efficient tools in business and management modeling for financial researchers and decision-makers in the Islamic banking sector. Originality/value: According to the author's knowledge, this paper is the first that provides data mining techniques for predicting the financial performance of the Islamic banking sector in Indonesia. [ABSTRACT FROM AUTHOR]

Big data technology has been developed rapidly in recent years. The performance improvement mechanism of targeted poverty alleviation is studied through the big data technology to further promote the comprehensive application of big data technology in poverty alleviation and development. Using the data mining knowledge to accurately identify the poor population under the framework of big data, compared with the traditional identification method, it is obviously more accurate and persuasive, which is also helpful to find out the real causes of poverty and assist the poor residents in the future. In the current targeted poverty alleviation work, the identification of poor households and the matching of assistance measures are mainly through the visiting of village cadres and the establishment of documents. Traditional methods are time-consuming, laborious, and difficult to manage. It always omits lots of useful family information. Therefore, new technologies need to be introduced to realize intelligent identification of poverty-stricken households and reduce labor costs. In this paper, we introduce a novel DBSCAN clustering algorithm via the edge computing-based deep neural network model for targeted poverty alleviation. First, we deploy an edge computing-based deep neural network model. Then, in this constructed model, we execute data mining for the poverty-stricken family. In this paper, the DBSCAN clustering algorithm is used to excavate the poverty features of the poor households and complete the intelligent identification of the poor households. In view of the current situation of high-dimensional and large-volume poverty alleviation data, the algorithm uses the relative density difference of grid to divide the data space into regions with different densities and adopts the DBSCAN algorithm to cluster the above result, which improves the accuracy of DBSCAN. This avoids the need for DBSCAN to traverse all data when searching for density connections. Finally, the proposed method is utilized for analyzing and mining the poverty alleviation data. The average accuracy is more than 96%. The average F -measure, NMI, and PRE values exceed 90%. The results show that it provides decision support for precise matching and intelligent pairing of village cadres in poverty alleviation work. [ABSTRACT FROM AUTHOR]

Modern data mining techniques using machine learning (ML) and deep learning (DL) algorithms have been shown to excel in the regression-based task of materials property prediction using various materials representations. In an attempt to improve the predictive performance of the deep neural network model, researchers have tried to add more layers as well as develop new architectural components to create sophisticated and deep neural network models that can aid in the training process and improve the predictive ability of the final model. However, usually, these modifications require a lot of computational resources, thereby further increasing the already large model training time, which is often not feasible, thereby limiting usage for most researchers. In this paper, we study and propose a deep neural network framework for regression-based problems comprising of fully connected layers that can work with any numerical vector-based materials representations as model input. We present a novel deep regression neural network, iBRNet, with branched skip connections and multiple schedulers, which can reduce the number of parameters used to construct the model, improve the accuracy, and decrease the training time of the predictive model. We perform the model training using composition-based numerical vectors representing the elemental fractions of the respective materials and compare their performance against other traditional ML and several known DL architectures. Using multiple datasets with varying data sizes for training and testing, We show that the proposed iBRNet models outperform the state-of-the-art ML and DL models for all data sizes. We also show that the branched structure and usage of multiple schedulers lead to fewer parameters and faster model training time with better convergence than other neural networks. Scientific contribution: The combination of multiple callback functions in deep neural networks minimizes training time and maximizes accuracy in a controlled computational environment with parametric constraints for the task of materials property prediction. [ABSTRACT FROM AUTHOR]

Multimedia data mining plays a crucial role in various fields, such as image and video analysis, natural language processing, and recommendation systems. Multimedia data refers to any form of data that involves multiple modes of communication, such as text, images, audio, and video. To effectively mine valuable insights from multimedia data, a new framework is proposed in this paper that employs a transformer-based intelligent deep neural network (DNN) model architecture. The framework includes an extensive data preprocessing step that involves obtaining multimedia data from internet searches and removing duplicates to ensure that each image is unique. The proposed transformer-based intelligent DNN model architecture processes the multimedia data in a hierarchical manner and utilizes shifted windows to achieve high accuracy in image classification task. The exploited dataset details are provided in the experimental evaluation section. Experimental results show that the proposed framework outperforms existing multimedia data mining methods in terms of accuracy and efficiency. This framework provides valuable insights that can be used in various applications, including content-based image retrieval, sentiment analysis, and automated captioning. [ABSTRACT FROM AUTHOR]

Understanding students' learning of knowledge components (KCs) is an important educational data mining task and enables many educational applications. However, in the domain of computing education, where program exercises require students to practice many KCs simultaneously, it is a challenge to attribute their errors to specific KCs and, therefore, to model student knowledge of these KCs. In this paper, we define this task as the KC attribution problem. We first demonstrate a novel approach to addressing this task using deep neural networks and explore its performance in identifying expert-defined KCs (RQ1). Because the labeling process takes costly expert resources, we further evaluate the effectiveness of transfer learning for KC attribution, using more easily acquired labels, such as problem correctness (RQ2). Finally, because prior research indicates the incorporation of educational theory in deep learning models could potentially enhance model performance, we investigated how to incorporate learning curves in the model design and evaluated their performance (RQ3). Our results show that in a supervised learning scenario, we can use a deep learning model, code2vec, to attribute KCs with a relatively high performance (AUC > 75% in two of the three examined KCs). Further using transfer learning, we achieve reasonable performance on the task without any costly expert labeling. However, the incorporation of learning curves shows limited effectiveness in this task. Our research lays important groundwork for personalized feedback for students based on which KCs they applied correctly, as well as more interpretable and accurate student models. [ABSTRACT FROM AUTHOR]

This paper addresses data mining and neural network model construction and analysis to design a data interaction process model based on data mining and topology visualization. This paper performs preprocessing data operations such as data filtering and cleaning of the collected data. A typical multichannel convolutional neural network (MCNN) in deep learning techniques is applied to alert students' academic performance. In addition, the network topology of the CNN is optimized to improve the performance of the model. The CNN has many hyperparameters that need to be tuned to construct an optimal model that can effectively interact with the data. In this paper, we propose a method to visualize the network topology within unstable regions to address the current problem of lacking an effective way to layout the network topology into specified areas. The technique transforms the network topology layout problem within the unstable region into a circular topology diffusion problem within a convex polygon, ensuring a clear, logical topology connection, and dramatically reducing the gaps in the area, making the layout more uniform beautiful. This paper constructs a real-time data interaction model based on JSON format and database triggers using message queues for reliable delivery. A platform-based real-time data interaction solution is designed by combining the timer method with the original key. The solution designed in this paper considers the real-time accuracy, security and reliability of data interaction. It satisfies the platform's initial and newly discovered requirements for data interaction. [ABSTRACT FROM AUTHOR]

Event extraction technology is one of the important researches in the field of information extraction, which helps people accurately retrieve, find, classify, and summarize effective information from a large amount of information streams. This paper uses the neural network hybrid model to identify the trigger words and event categories of the legal domain knowledge graph events, extracts the events of interest from a large amount of free text, and displays them in a structured format. First, the original text is preprocessed, and then, the distributed semantic word vector is combined with the dependent syntactic structure and location attributes to create a semantic representation in the form of a vector. The combined deep learning model is used to extract activated words, the long-term memory loop neural network uses temporal semantics to extract deep features, and the convergent neural network completes the extraction of activated words and event categories. Finally, the experimental results show that the accuracy of event extraction of the neural network hybrid model designed in this paper has reached 77.1%, and the recall rate has reached 76.8%, which is greatly improved compared with the traditional model. [ABSTRACT FROM AUTHOR]

This paper proposes a network model recurrent fully connected network (RFC-Net) based on recurrent full convolution and polarization change. RFC-Net enriches the network by reconstructing and fine-tuning the fully convolutional network and adding recurrent convolutions to it. By studying the data mining technology of multidimensional association rules, based on the existing algorithms, this paper improves the shortcomings of the algorithms and realizes an efficient and practical method for data mining based on interdimensional multidimensional association rules. On the basis of mastering the actual student information, the effectiveness of the method is tested, and an employment analysis system based on association rules is established. Aiming at the fact that traditional grade prediction methods ignore the different influences of different behavioral characteristics on grades, and considering that behavioral data in different periods have different influences on student grades, the grade prediction problem is abstracted into a time series classification problem. The mechanism is combined with long short-term memory neural network to construct a performance prediction model based on Attention-BiLSTM. Experiments show that the prediction model proposed in this paper improves the accuracy and effectively improves the prediction quality compared with the logistic regression model with a better prediction effect in the traditional benchmark model and the long short-term memory neural network model without the introduction of the attention mechanism. Research shows that physical performance and academic performance are not contradictory. We must face up to the status of physical exercise in schools; as long as physical exercise is properly arranged, it can inspire students to form a spirit of unity, interaction, positivity, and perseverance in cultural studies. [ABSTRACT FROM AUTHOR]

With the development of deep convolutional neural network, the performance of instance segmentation algorithms has been significantly improved. However, it is difficult to implement quick and accurate near-shore ship segmentation due to the complex background and arbitrary ship orientation, which makes segmentation challenging. To improve the segmentation efficiency of near-shore ships, this paper proposes a ship segmentation network based on Polar Mask, named I-Polar Mask. Specifically, we construct a Specific Polar Template (SPT) and Oriented Polar IoU (Intersection over Union) to better match the ship contour. Furthermore, oriented polar centre-ness was designed to reduce the weight of low-quality masks more effectively. In addition, a Context Information Extraction Module (CIEM) is built to reduce the influence of complex backgrounds and make the segmentation more accurate. To verify the effectiveness of the proposed algorithm, we collected 1015 images of near-shore areas from Google Maps and labelled them with Labelme to construct a ship instance segmentation dataset, called I-Ship. Extensive experiments on I-Ship show that the AP value of I-Polar Mask is improved by 5.9% compared with Polar Mask, which is a significant improvement. Compared with advanced methods, I-Polar Mask outperforms both quantitative and qualitative aspects. [ABSTRACT FROM AUTHOR]

With the vigorousness of the knitting industry, defect detection and classification of weft-knitted fabrics have become the research fields with extensive application value. However, convolution neural network models suffer from the limitation of convolutional operation, which makes them unable to capture the global features of fabric images abundantly. Although the transformer can compensate for this deficiency, it still has shortcomings such as poor small target recognition and unsatisfactory local information extraction ability. In order sufficiently to actualize the mutual support of relative advantages between the convolution neural network and the transformer, a Swin transformer deformable convolutional network integrated model is proposed in this paper. The Swin transformer deformable convolutional network utilizes the self-attention mechanism with global perception to establish dependencies comprehensively between long-range elements. Meanwhile, the deformable convolution is introduced according to the shaped character)istics of defects to extract local features effectively. Furthermore, a dataset containing 5474 images of weft-knitted fabrics was designed due to the less adequate databases. Experimental results on our weft-knitted fabric dataset and the Irish Longitudinal Study on Ageing (TILDA) database demonstrated that the proposed Swin transformer deformable convolutional network is superior to current state-of-the-art methods and has immense potential in fabric defect detection and classification. [ABSTRACT FROM AUTHOR]

Purpose: Taiwan experiences frequent seismic activity. Major earthquakes in recent history have seriously damaged the school buildings. School buildings in Taiwan are intended to serve both as places of education and as temporary shelters in the aftermath of major earthquakes. Therefore, the seismic performance assessments of school buildings are critical issues that deserve investigation. Design/methodology/approach: This paper develops a methodology that uses principal component analysis to generalize the seismic factors from the basic seismic parameters of school buildings, uses data mining to cluster different school building sizes and uses grey theory to analyze the relationship between seismic factors and the seismic performance of school buildings. Additionally, this paper employs the Artificial Neural Network (ANN) to deduce the seismic assessment model for school buildings. Finally, it adopts support vector machine to validate the ANN's deductive results. Findings: An empirical study was conducted on 326 school buildings in the central area of Taichung City, Taiwan, to illustrate the effectiveness of the proposed approach. Results show that thickness of wall and width of middle-row column relate significantly with school-building seismic performance. Originality/value: This paper provides a model that structural engineers or architects may use to design school buildings that are adequately resistant to earthquakes as well as a reference for future academic research. [ABSTRACT FROM AUTHOR]

Combining computational technologies and humanities is an ongoing effort aimed at making resources such as texts, images, audio, video, and other artifacts digitally available, searchable, and analyzable. In recent years, deep neural networks (DNN) dominate the field of automatic text analysis and natural language processing (NLP), in some cases presenting a super‐human performance. DNNs are the state‐of‐the‐art machine learning algorithms solving many NLP tasks that are relevant for Digital Humanities (DH) research, such as spell checking, language detection, entity extraction, author detection, question answering, and other tasks. These supervised algorithms learn patterns from a large number of "right" and "wrong" examples and apply them to new examples. However, using DNNs for analyzing the text resources in DH research presents two main challenges: (un)availability of training data and a need for domain adaptation. This paper explores these challenges by analyzing multiple use‐cases of DH studies in recent literature and their possible solutions and lays out a practical decision model for DH experts for when and how to choose the appropriate deep learning approaches for their research. Moreover, in this paper, we aim to raise awareness of the benefits of utilizing deep learning models in the DH community. [ABSTRACT FROM AUTHOR]

Simple Summary: In radiotherapy treatment, organs at risk and target volumes are contoured by the clinicians to prepare a dosimetry plan. In retrospective data, these structures are not often standardised to universal names across the patients plans, which is required to enable data mining and analysis. In this paper, a new method was proposed and evaluated to automatically standardise radiotherapy structures names using machine learning algorithms. The proposed approach was deployed over a dataset with 1613 patients collected from Liverpool & Macarthur Cancer Therapy Centres, New South Wales, Australia. It was concluded that machine learning techniques can standardise the dosimetry plan structures, taking into consideration the integration of multiple modalities representing each structure during the training process. In progressing the use of big data in health systems, standardised nomenclature is required to enable data pooling and analyses. In many radiotherapy planning systems and their data archives, target volumes (TV) and organ-at-risk (OAR) structure nomenclature has not been standardised. Machine learning (ML) has been utilised to standardise volumes nomenclature in retrospective datasets. However, only subsets of the structures have been targeted. Within this paper, we proposed a new approach for standardising all the structures nomenclature by using multi-modal artificial neural networks. A cohort consisting of 1613 breast cancer patients treated with radiotherapy was identified from Liverpool & Macarthur Cancer Therapy Centres, NSW, Australia. Four types of volume characteristics were generated to represent each target and OAR volume: textual features, geometric features, dosimetry features, and imaging data. Five datasets were created from the original cohort, the first four represented different subsets of volumes and the last one represented the whole list of volumes. For each dataset, 15 sets of combinations of features were generated to investigate the effect of using different characteristics on the standardisation performance. The best model reported 99.416% classification accuracy over the hold-out sample when used to standardise all the nomenclatures in a breast cancer radiotherapy plan into 21 classes. Our results showed that ML based automation methods can be used for standardising naming conventions in a radiotherapy plan taking into consideration the inclusion of multiple modalities to better represent each volume. [ABSTRACT FROM AUTHOR]

This paper represents a neural network model for the Web page information extraction based on the depth learning technology, and implements the model algorithm using the TensorFlow system. We then complete a detailed experimental analysis of the information extraction effect of Web pages on the same website, then show statistics on the accuracy index of the page information extraction, and optimize some parameters in the model according to the experimental results. On the premise of achieving ideal experimental results, an algorithm for migrating the model to the same pages of other websites for information extraction is proposed, and the experimental results are analyzed. Although the overall effect of the experiment is not as good as that of the page information extraction in different websites, it is far more effective than that of using the model directly on new websites. A new method is proposed to improve the portability of the information extraction system based on machine learning technology. At the same time, the deep nonlinear learning method of the depth learning model can prove deeper features, can have a more essential description of the abstract language, and can better express and understand sentences from the syntactic and semantic levels. [ABSTRACT FROM AUTHOR]

Data mining plays a crucial role in extracting meaningful knowledge fromlarge-scale data repositories, such as data warehouses and databases. Association rule mining, a fundamental process in data mining, involves discovering correlations, patterns, and causal structures within datasets. In the healthcare domain, association rules offer valuable opportunities for building knowledge bases, enabling intelligent diagnoses, and extracting invaluable information rapidly. This paper presents a novel approach called the Machine Learning based Association Rule Mining and Classification for Healthcare Data Management System (MLARMC-HDMS). The MLARMC-HDMS technique integrates classification and association rule mining (ARM) processes. Initially, the chimp optimization algorithm-based feature selection (COAFS) technique is employed within MLARMC-HDMS to select relevant attributes. Inspired by the foraging behavior of chimpanzees, the COA algorithm mimics their search strategy for food. Subsequently, the classification process utilizes stochastic gradient descent with a multilayer perceptron (SGD-MLP) model, while the Apriori algorithm determines attribute relationships.We propose a COA-based feature selection approach for medical data classification using machine learning techniques. This approach involves selecting pertinent features from medical datasets through COA and training machine learning models using the reduced feature set. We evaluate the performance of our approach on various medical datasets employing diverse machine learning classifiers. Experimental results demonstrate that our proposed approach surpasses alternative feature selection methods, achieving higher accuracy and precision rates in medical data classification tasks. The study showcases the effectiveness and efficiency of the COA-based feature selection approach in identifying relevant features, thereby enhancing the diagnosis and treatment of various diseases. To provide further validation, we conduct detailed experiments on a benchmark medical dataset, revealing the superiority of the MLARMCHDMS model over other methods, with a maximum accuracy of 99.75%. Therefore, this research contributes to the advancement of feature selection techniques in medical data classification and highlights the potential for improving healthcare outcomes through accurate and efficient data analysis. The presented MLARMC-HDMS framework and COA-based feature selection approach offer valuable insights for researchers and practitioners working in the field of healthcare data mining and machine learning. [ABSTRACT FROM AUTHOR]

Data mining with cloud computing is the process of extracting knowledge and ideas from large data sets using the resources and capabilities of cloud computing platforms. Cloud computing provides a scalable and costeffective way to store, process and analyze large data sets. Cloud mining refers to the combination of dynamic, distributed real-time data management technology and data mining technology to achieve dynamic, distributed and efficient real-time processing, extraction and analysis of large amounts of data. In this paper, we present a datamining platform based on cloud computing. [ABSTRACT FROM AUTHOR]

Physiological signal-based biometrics are gaining increasing attention in the context of increasing privacy and security requirements. This paper proposes a novel electrocardiogram (ECG)-based algorithm to be used for human identification by integrating multiple local feature vectors with sparse-constraint-based sparse coding (SCSC) and bidirectional long short-term memory (BLSTM). Three local feature vectors of ECG signals: morphology characteristics in the time domain, instantaneous characteristics in the frequency domain, and phase spectral characteristics in the phase domain are constructed. Sparsity constraints to model this relationship are imposed because ECGs show high inter-class similarity and subtle intra-class differences in these three domains, and traditional sparse coding (SC) can only learn from a single dictionary. This paper joints optimization of the summed reconstruction error, the sparsity constraints of the correlations and the differences between the feature vectors, proposed the SCSC algorithm. Via this approach, the overlap problem of local feature vectors is solved and a lightweight and interpretable feature vector is obtained. Additionally, the BLSTM-based deep neural network model is supplemented for exploring the spatial information of the reconstructed feature vectors, and a more representative and discriminative signal feature representation is obtained. Comparing five classical machine learning and deep learning algorithms within 360 public samples, using two protocols, we show that, in addition to multiscale information extraction, joint encoding of the correlations and differences between local feature vectors is critically important for feature discrimination. The experimental results demonstrated a high identification accuracy of 99.44%, indicating that the proposed algorithm has practical utility in network information security. [ABSTRACT FROM AUTHOR]

Purpose: This paper presents a new approach of productivity estimation of cutter suction dredger operation through data mining and learning from real-time big data. Design/methodology/approach: The paper used big data, data mining and machine learning techniques to extract features of cutter suction dredgers (CSD) for predicting its productivity. ElasticNet-SVR (Elastic Net-Support Vector Machine) method is used to filter the original monitoring data. Along with the actual working conditions of CSD, 15 features were selected. Then, a box plot was used to clean the corresponding data by filtering out outliers. Finally, four algorithms, namely SVR (Support Vector Regression), XGBoost (Extreme Gradient Boosting), LSTM (Long-Short Term Memory Network) and BP (Back Propagation) Neural Network, were used for modeling and testing. Findings: The paper provided a comprehensive forecasting framework for productivity estimation including feature selection, data processing and model evaluation. The optimal coefficient of determination (R2) of four algorithms were all above 80.0%, indicating that the features selected were representative. Finally, the BP neural network model coupled with the SVR model was selected as the final model. Originality/value: Machine-learning algorithm incorporating domain expert judgments was used to select predictive features. The final optimal coefficient of determination (R2) of the coupled model of BP neural network and SVR is 87.6%, indicating that the method proposed in this paper is effective for CSD productivity estimation. [ABSTRACT FROM AUTHOR]

While recent developments of neural network (NN) models have led to a series of record-breaking achievements in many applications, the lack of sufficiently good datasets remains a problem for some applications. For such a problem, we can however exploit a large number of unstructured text corpora as an external knowledge to complement the training data, and most prevailing neural network solutions employ word embedding methods for such purposes. In this paper, we propose LDA-Reg, a novel knowledge driven regularization framework based on Latent Dirichlet Allocation (LDA) as an alternative to the word embedding methods to adaptively utilize abundant external knowledge and to interpret the NN model. For the joint learning of the parameters, we propose EM-SGD, an effective update method which incorporates Expectation Maximization (EM) and Stochastic Gradient Descent (SGD) to update parameters iteratively. Moreover, we also devise a lazy update and sparse update method for the high-dimensional inputs and sparse inputs respectively. We validate the effectiveness of our regularization framework through an extensive experimental study over real world and standard benchmark datasets. The results show that our proposed framework not only achieves significant improvement over state-of-the-art word embedding methods but also learns interpretable and significant topics for various tasks. [ABSTRACT FROM AUTHOR]

Pedestrians are vulnerable road users subject to severer injuries and higher fatality risk in motor vehicle crashes due to limited protection. An important portion of vehicle–pedestrian crashes occurred at intersections due to the complex movements of various types of road users and the conflicts among them. To address the safety concern, this paper investigates the contributing factors to the severity of vehicle–pedestrian crashes at intersections based on a 3-year crash dataset of Hong Kong. For the crash severity modelling process, the crash dataset is mass and complicated. To tackle the class imbalance issue of the crash severity level, data resampling method is firstly applied. Then, various data mining algorithms, namely, classification and regression tree (CART) model, gradient boosting (GB) model, random forest (RF) model, artificial neural network (ANN) model and support vector machine (SVM) model, have been applied. The performance of these models have also been compared with the logistic regression model commonly applied in the literature. The ANN model which has the best performance is selected to determine the most significant contributing factors to the fatal and severe crashes, and the marginal effects of these factors are also analysed. Results show that the likelihood of fatal and severe vehicle–pedestrian crashes at intersections increase when there is light rain and where the junction control type is traffic signal and no control. On the other hand, the crash severity tends to decrease when the weather condition is clear, the light condition is daylight and dark, and in the districts of Kwun Tong, Kowloon City, Central and Western, and Sham Shui Po. Based on the results, policy implications and counter-measures on reducing the fatal and severe vehicle–pedestrian crashes at intersections have been recommended. [ABSTRACT FROM AUTHOR]

Deep forest methods have gradually emerged as a well-liked substitute for conventional deep neural networks in diagnosing faults in mechanical systems. However, in practical industrial applications, limited training data and severe noise interference pose significant challenges to these models. Existing deep forest models have limitations in information extraction, making it difficult to handle the complexities of industrial environments. To better meet the practical needs of industrial applications, this paper proposes an improved deep forest model – sgic-Forest, specifically designed for fault diagnosis of small sample gearboxes in noisy environments. First, we developed a stacked multi-grained scanning module, which enhances the diversity of feature extraction by integrating the advantages of multiple base learners, thereby better addressing the complexities of industrial data. Secondly, we introduced an important feature selection module, which effectively filters out irrelevant information, significantly improving the model’s robustness in high-noise environments. Experiments on two gearbox datasets show that the proposed method outperforms the basic deep forest model and mainstream deep learning methods in terms of diagnostic accuracy under small sample and noisy conditions. [ABSTRACT FROM AUTHOR]