María Novo-Lourés, Reyes Pavón, Rosalía Laza, José R. Méndez, David Ruano-Ordás
With the advent and improvement of ontological dictionaries (WordNet, Babelnet), the use of synsets-based text representations is gaining popularity in classification tasks. More recently, ontological dictionaries were used for reducing dimensionality in this kind of representation (e.g., Semantic Dimensionality Reduction System (SDRS) (Vélez de Mendizabal et al., 2020)). These approaches are based on the combination of semantically related columns by taking advantage of semantic information extracted from ontological dictionaries. Their main advantage is that they not only eliminate features but can also combine them, minimizing (low-loss) or avoiding (lossless) the loss of information. The most recent (and accurate) techniques included in this group are based on using evolutionary algorithms to find how many features can be grouped to reduce false positive (FP) and false negative (FN) errors obtained. The main limitation of these evolutionary-based schemes is the computational requirements derived from the use of optimization algorithms. The contribution of this study is a new lossless feature reduction scheme exploiting information from ontological dictionaries, which achieves slightly better accuracy (specially in FP errors) than optimization-based approaches but using far fewer computational resources. Instead of using computationally expensive evolutionary algorithms, our proposal determines whether two columns (synsets) can be combined by observing whether the instances included in a dataset (e.g., training dataset) containing these synsets are mostly of the same class. The study includes experiments using three datasets and a detailed comparison with two previous optimization-based approaches.
随着本体字典（WordNet、Babelnet）的出现和改进，基于同义词集的文本表示法在分类任务中越来越受欢迎。最近，本体词典被用于降低此类表示的维度（例如语义维度降低系统（SDRS）（Vélez de Mendizabal 等人，2020 年））。这些方法的基础是利用从本体字典中提取的语义信息，将语义相关的列进行组合。它们的主要优势在于不仅能消除特征，还能组合特征，最大限度地减少（低损耗）或避免（无损耗）信息损失。这类技术中最新（也是最准确）的技术是基于进化算法，找出有多少特征可以通过分组来减少假阳性（FP）和假阴性（FN）误差。这些基于进化算法的方案的主要局限性在于使用优化算法所带来的计算要求。本研究的贡献在于利用本体字典中的信息，提出了一种新的无损特征缩减方案，与基于优化的方法相比，该方案的准确率（尤其是在 FP 错误方面）略高，但使用的计算资源却要少得多。我们的方案不使用计算成本高昂的进化算法，而是通过观察包含这些同义词集的数据集（如训练数据集）中的实例是否大多属于同一类别，来确定是否可以合并两列（同义词集）。研究包括使用三个数据集进行的实验，以及与之前两种基于优化的方法进行的详细比较。
{"title":"An enhanced algorithm for semantic-based feature reduction in spam filtering","authors":"María Novo-Lourés, Reyes Pavón, Rosalía Laza, José R. Méndez, David Ruano-Ordás","doi":"10.7717/peerj-cs.2206","DOIUrl":"https://doi.org/10.7717/peerj-cs.2206","url":null,"abstract":"With the advent and improvement of ontological dictionaries (WordNet, Babelnet), the use of synsets-based text representations is gaining popularity in classification tasks. More recently, ontological dictionaries were used for reducing dimensionality in this kind of representation (e.g., Semantic Dimensionality Reduction System (SDRS) (Vélez de Mendizabal et al., 2020)). These approaches are based on the combination of semantically related columns by taking advantage of semantic information extracted from ontological dictionaries. Their main advantage is that they not only eliminate features but can also combine them, minimizing (low-loss) or avoiding (lossless) the loss of information. The most recent (and accurate) techniques included in this group are based on using evolutionary algorithms to find how many features can be grouped to reduce false positive (FP) and false negative (FN) errors obtained. The main limitation of these evolutionary-based schemes is the computational requirements derived from the use of optimization algorithms. The contribution of this study is a new lossless feature reduction scheme exploiting information from ontological dictionaries, which achieves slightly better accuracy (specially in FP errors) than optimization-based approaches but using far fewer computational resources. Instead of using computationally expensive evolutionary algorithms, our proposal determines whether two columns (synsets) can be combined by observing whether the instances included in a dataset (e.g., training dataset) containing these synsets are mostly of the same class. The study includes experiments using three datasets and a detailed comparison with two previous optimization-based approaches.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"46 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141867286","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Personalized learning resource recommendations may help resolve the difficulties of online education that include learning mazes and information overload. However, existing personalized learning resource recommendation algorithms have shortcomings such as low accuracy and low efficiency. This study proposes a deep recommendation system algorithm based on a knowledge graph (D-KGR) that includes four data processing units. These units are the recommendation unit (RS unit), the knowledge graph feature representation unit (KGE unit), the cross compression unit (CC unit), and the feature extraction unit (FE unit). This model integrates technologies including the knowledge graph, deep learning, neural network, and data mining. It introduces cross compression in the feature learning process of the knowledge graph and predicts user attributes. Multimodal technology is used to optimize the process of project attribute processing; text type attributes, multivalued type attributes, and other type attributes are processed separately to reconstruct the knowledge graph. A convolutional neural network algorithm is introduced in the reconstruction process to optimize the data feature qualities. Experimental analysis was conducted from two aspects of algorithm efficiency and accuracy, and the particle swarm optimization, neural network, and knowledge graph algorithms were compared. Several tests showed that the deep recommendation system algorithm had obvious advantages when the number of learning resources and users exceeded 1,000. It has the ability to integrate systems such as the particle swarm optimization iterative classification, neural network intelligent simulation, and low resource consumption. It can quickly process massive amounts of information data, reduce algorithm complexity and requires less time and had lower costs. Our algorithm also has better efficiency and accuracy.
{"title":"A knowledge graph algorithm enabled deep recommendation system","authors":"Yan Wang, Xiao Feng Ma, Miao Zhu","doi":"10.7717/peerj-cs.2010","DOIUrl":"https://doi.org/10.7717/peerj-cs.2010","url":null,"abstract":"Personalized learning resource recommendations may help resolve the difficulties of online education that include learning mazes and information overload. However, existing personalized learning resource recommendation algorithms have shortcomings such as low accuracy and low efficiency. This study proposes a deep recommendation system algorithm based on a knowledge graph (D-KGR) that includes four data processing units. These units are the recommendation unit (RS unit), the knowledge graph feature representation unit (KGE unit), the cross compression unit (CC unit), and the feature extraction unit (FE unit). This model integrates technologies including the knowledge graph, deep learning, neural network, and data mining. It introduces cross compression in the feature learning process of the knowledge graph and predicts user attributes. Multimodal technology is used to optimize the process of project attribute processing; text type attributes, multivalued type attributes, and other type attributes are processed separately to reconstruct the knowledge graph. A convolutional neural network algorithm is introduced in the reconstruction process to optimize the data feature qualities. Experimental analysis was conducted from two aspects of algorithm efficiency and accuracy, and the particle swarm optimization, neural network, and knowledge graph algorithms were compared. Several tests showed that the deep recommendation system algorithm had obvious advantages when the number of learning resources and users exceeded 1,000. It has the ability to integrate systems such as the particle swarm optimization iterative classification, neural network intelligent simulation, and low resource consumption. It can quickly process massive amounts of information data, reduce algorithm complexity and requires less time and had lower costs. Our algorithm also has better efficiency and accuracy.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"20 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141873064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The SARS-CoV-2 virus, which induces an acute respiratory illness commonly referred to as COVID-19, had been designated as a pandemic by the World Health Organization due to its highly infectious nature and the associated public health risks it poses globally. Identifying the critical factors for predicting mortality is essential for improving patient therapy. Unlike other data types, such as computed tomography scans, x-radiation, and ultrasounds, basic blood test results are widely accessible and can aid in predicting mortality. The present research advocates the utilization of machine learning (ML) methodologies for predicting the likelihood of infectious disease like COVID-19 mortality by leveraging blood test data. Age, LDH (lactate dehydrogenase), lymphocytes, neutrophils, and hs-CRP (high-sensitivity C-reactive protein) are five extremely potent characteristics that, when combined, can accurately predict mortality in 96% of cases. By combining XGBoost feature importance with neural network classification, the optimal approach can predict mortality with exceptional accuracy from infectious disease, along with achieving a precision rate of 90% up to 16 days before the event. The studies suggested model’s excellent predictive performance and practicality were confirmed through testing with three instances that depended on the days to the outcome. By carefully analyzing and identifying patterns in these significant biomarkers insightful information has been obtained for simple application. This study offers potential remedies that could accelerate decision-making for targeted medical treatments within healthcare systems, utilizing a timely, accurate, and reliable method.
{"title":"Real-time infectious disease endurance indicator system for scientific decisions using machine learning and rapid data processing","authors":"Shivendra Dubey, Dinesh Kumar Verma, Mahesh Kumar","doi":"10.7717/peerj-cs.2062","DOIUrl":"https://doi.org/10.7717/peerj-cs.2062","url":null,"abstract":"The SARS-CoV-2 virus, which induces an acute respiratory illness commonly referred to as COVID-19, had been designated as a pandemic by the World Health Organization due to its highly infectious nature and the associated public health risks it poses globally. Identifying the critical factors for predicting mortality is essential for improving patient therapy. Unlike other data types, such as computed tomography scans, x-radiation, and ultrasounds, basic blood test results are widely accessible and can aid in predicting mortality. The present research advocates the utilization of machine learning (ML) methodologies for predicting the likelihood of infectious disease like COVID-19 mortality by leveraging blood test data. Age, LDH (lactate dehydrogenase), lymphocytes, neutrophils, and hs-CRP (high-sensitivity C-reactive protein) are five extremely potent characteristics that, when combined, can accurately predict mortality in 96% of cases. By combining XGBoost feature importance with neural network classification, the optimal approach can predict mortality with exceptional accuracy from infectious disease, along with achieving a precision rate of 90% up to 16 days before the event. The studies suggested model’s excellent predictive performance and practicality were confirmed through testing with three instances that depended on the days to the outcome. By carefully analyzing and identifying patterns in these significant biomarkers insightful information has been obtained for simple application. This study offers potential remedies that could accelerate decision-making for targeted medical treatments within healthcare systems, utilizing a timely, accurate, and reliable method.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"61 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141867402","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yan Dong, Yundong Liu, Yuhua Cheng, Guangshuai Gao, Kai Chen, Chunlei Li
Accurate localization of objects of interest in remote sensing images (RSIs) is of great significance for object identification, resource management, decision-making and disaster relief response. However, many difficulties, like complex backgrounds, dense target quantities, large-scale variations, and small-scale objects, which make the detection accuracy unsatisfactory. To improve the detection accuracy, we propose an Adaptive Adjacent Context Negotiation Network (A2CN-Net). Firstly, the composite fast Fourier convolution (CFFC) module is given to reduce the information loss of small objects, which is inserted into the backbone network to obtain spectral global context information. Then, the Global Context Information Enhancement (GCIE) module is given to capture and aggregate global spatial features, which is beneficial for locating objects of different scales. Furthermore, to alleviate the aliasing effect caused by the fusion of adjacent feature layers, a novel Adaptive Adjacent Context Negotiation network (A2CN) is given to adaptive integration of multi-level features, which consists of local and adjacent branches, with the local branch adaptively highlighting feature information and the adjacent branch introducing global information at the adjacent level to enhance feature representation. In the meantime, considering the variability in the focus of feature layers in different dimensions, learnable weights are applied to the local and adjacent branches for adaptive feature fusion. Finally, extensive experiments are performed in several available public datasets, including DIOR and DOTA-v1.0. Experimental studies show that A2CN-Net can significantly boost detection performance, with mAP increasing to 74.2% and 79.2%, respectively.
{"title":"Adaptive adjacent context negotiation network for object detection in remote sensing imagery","authors":"Yan Dong, Yundong Liu, Yuhua Cheng, Guangshuai Gao, Kai Chen, Chunlei Li","doi":"10.7717/peerj-cs.2199","DOIUrl":"https://doi.org/10.7717/peerj-cs.2199","url":null,"abstract":"Accurate localization of objects of interest in remote sensing images (RSIs) is of great significance for object identification, resource management, decision-making and disaster relief response. However, many difficulties, like complex backgrounds, dense target quantities, large-scale variations, and small-scale objects, which make the detection accuracy unsatisfactory. To improve the detection accuracy, we propose an Adaptive Adjacent Context Negotiation Network (A2CN-Net). Firstly, the composite fast Fourier convolution (CFFC) module is given to reduce the information loss of small objects, which is inserted into the backbone network to obtain spectral global context information. Then, the Global Context Information Enhancement (GCIE) module is given to capture and aggregate global spatial features, which is beneficial for locating objects of different scales. Furthermore, to alleviate the aliasing effect caused by the fusion of adjacent feature layers, a novel Adaptive Adjacent Context Negotiation network (A2CN) is given to adaptive integration of multi-level features, which consists of local and adjacent branches, with the local branch adaptively highlighting feature information and the adjacent branch introducing global information at the adjacent level to enhance feature representation. In the meantime, considering the variability in the focus of feature layers in different dimensions, learnable weights are applied to the local and adjacent branches for adaptive feature fusion. Finally, extensive experiments are performed in several available public datasets, including DIOR and DOTA-v1.0. Experimental studies show that A2CN-Net can significantly boost detection performance, with mAP increasing to 74.2% and 79.2%, respectively.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"156 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141867367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Image segmentation is a crucial process in the field of image processing. Multilevel threshold segmentation is an effective image segmentation method, where an image is segmented into different regions based on multilevel thresholds for information analysis. However, the complexity of multilevel thresholding increases dramatically as the number of thresholds increases. To address this challenge, this article proposes a novel hybrid algorithm, termed differential evolution-golden jackal optimizer (DEGJO), for multilevel thresholding image segmentation using the minimum cross-entropy (MCE) as a fitness function. The DE algorithm is combined with the GJO algorithm for iterative updating of position, which enhances the search capacity of the GJO algorithm. The performance of the DEGJO algorithm is assessed on the CEC2021 benchmark function and compared with state-of-the-art optimization algorithms. Additionally, the efficacy of the proposed algorithm is evaluated by performing multilevel segmentation experiments on benchmark images. The experimental results demonstrate that the DEGJO algorithm achieves superior performance in terms of fitness values compared to other metaheuristic algorithms. Moreover, it also yields good results in quantitative performance metrics such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and feature similarity index (FSIM) measurements.
{"title":"An efficient hybrid differential evolution-golden jackal optimization algorithm for multilevel thresholding image segmentation","authors":"Xianmeng Meng, Linglong Tan, Yueqin Wang","doi":"10.7717/peerj-cs.2121","DOIUrl":"https://doi.org/10.7717/peerj-cs.2121","url":null,"abstract":"Image segmentation is a crucial process in the field of image processing. Multilevel threshold segmentation is an effective image segmentation method, where an image is segmented into different regions based on multilevel thresholds for information analysis. However, the complexity of multilevel thresholding increases dramatically as the number of thresholds increases. To address this challenge, this article proposes a novel hybrid algorithm, termed differential evolution-golden jackal optimizer (DEGJO), for multilevel thresholding image segmentation using the minimum cross-entropy (MCE) as a fitness function. The DE algorithm is combined with the GJO algorithm for iterative updating of position, which enhances the search capacity of the GJO algorithm. The performance of the DEGJO algorithm is assessed on the CEC2021 benchmark function and compared with state-of-the-art optimization algorithms. Additionally, the efficacy of the proposed algorithm is evaluated by performing multilevel segmentation experiments on benchmark images. The experimental results demonstrate that the DEGJO algorithm achieves superior performance in terms of fitness values compared to other metaheuristic algorithms. Moreover, it also yields good results in quantitative performance metrics such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and feature similarity index (FSIM) measurements.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"86 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141867384","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As the pandemic continues to pose challenges to global public health, developing effective predictive models has become an urgent research topic. This study aims to explore the application of multi-objective optimization methods in selecting infectious disease prediction models and evaluate their impact on improving prediction accuracy, generalizability, and computational efficiency. In this study, the NSGA-II algorithm was used to compare models selected by multi-objective optimization with those selected by traditional single-objective optimization. The results indicate that decision tree (DT) and extreme gradient boosting regressor (XGBoost) models selected through multi-objective optimization methods outperform those selected by other methods in terms of accuracy, generalizability, and computational efficiency. Compared to the ridge regression model selected through single-objective optimization methods, the decision tree (DT) and XGBoost models demonstrate significantly lower root mean square error (RMSE) on real datasets. This finding highlights the potential advantages of multi-objective optimization in balancing multiple evaluation metrics. However, this study’s limitations suggest future research directions, including algorithm improvements, expanded evaluation metrics, and the use of more diverse datasets. The conclusions of this study emphasize the theoretical and practical significance of multi-objective optimization methods in public health decision support systems, indicating their wide-ranging potential applications in selecting predictive models.
{"title":"Enhancing infectious disease prediction model selection with multi-objective optimization: an empirical study","authors":"Deren Xu, Weng Howe Chan, Habibollah Haron","doi":"10.7717/peerj-cs.2217","DOIUrl":"https://doi.org/10.7717/peerj-cs.2217","url":null,"abstract":"As the pandemic continues to pose challenges to global public health, developing effective predictive models has become an urgent research topic. This study aims to explore the application of multi-objective optimization methods in selecting infectious disease prediction models and evaluate their impact on improving prediction accuracy, generalizability, and computational efficiency. In this study, the NSGA-II algorithm was used to compare models selected by multi-objective optimization with those selected by traditional single-objective optimization. The results indicate that decision tree (DT) and extreme gradient boosting regressor (XGBoost) models selected through multi-objective optimization methods outperform those selected by other methods in terms of accuracy, generalizability, and computational efficiency. Compared to the ridge regression model selected through single-objective optimization methods, the decision tree (DT) and XGBoost models demonstrate significantly lower root mean square error (RMSE) on real datasets. This finding highlights the potential advantages of multi-objective optimization in balancing multiple evaluation metrics. However, this study’s limitations suggest future research directions, including algorithm improvements, expanded evaluation metrics, and the use of more diverse datasets. The conclusions of this study emphasize the theoretical and practical significance of multi-objective optimization methods in public health decision support systems, indicating their wide-ranging potential applications in selecting predictive models.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"51 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141867366","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Classifying images is one of the most important tasks in computer vision. Recently, the best performance for image classification tasks has been shown by networks that are both deep and well-connected. These days, most datasets are made up of a fixed number of color images. The input images are taken in red green blue (RGB) format and classified without any changes being made to the original. It is observed that color spaces (basically changing original RGB images) have a major impact on classification accuracy, and we delve into the significance of color spaces. Moreover, datasets with a highly variable number of classes, such as the PlantVillage dataset utilizing a model that incorporates numerous color spaces inside the same model, achieve great levels of accuracy, and different classes of images are better represented in different color spaces. Furthermore, we demonstrate that this type of model, in which the input is preprocessed into many color spaces simultaneously, requires significantly fewer parameters to achieve high accuracy for classification. The proposed model basically takes an RGB image as input, turns it into seven separate color spaces at once, and then feeds each of those color spaces into its own Convolutional Neural Network (CNN) model. To lessen the load on the computer and the number of hyperparameters needed, we employ group convolutional layers in the proposed CNN model. We achieve substantial gains over the present state-of-the-art methods for the classification of crop disease.
{"title":"CropGCNN: color space-based crop disease classification using group convolutional neural network","authors":"Naeem Ahmad, Shubham Singh, Mohamed Fahad AlAjmi, Afzal Hussain, Khalid Raza","doi":"10.7717/peerj-cs.2136","DOIUrl":"https://doi.org/10.7717/peerj-cs.2136","url":null,"abstract":"Classifying images is one of the most important tasks in computer vision. Recently, the best performance for image classification tasks has been shown by networks that are both deep and well-connected. These days, most datasets are made up of a fixed number of color images. The input images are taken in red green blue (RGB) format and classified without any changes being made to the original. It is observed that color spaces (basically changing original RGB images) have a major impact on classification accuracy, and we delve into the significance of color spaces. Moreover, datasets with a highly variable number of classes, such as the PlantVillage dataset utilizing a model that incorporates numerous color spaces inside the same model, achieve great levels of accuracy, and different classes of images are better represented in different color spaces. Furthermore, we demonstrate that this type of model, in which the input is preprocessed into many color spaces simultaneously, requires significantly fewer parameters to achieve high accuracy for classification. The proposed model basically takes an RGB image as input, turns it into seven separate color spaces at once, and then feeds each of those color spaces into its own Convolutional Neural Network (CNN) model. To lessen the load on the computer and the number of hyperparameters needed, we employ group convolutional layers in the proposed CNN model. We achieve substantial gains over the present state-of-the-art methods for the classification of crop disease.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"361 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141867368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Layout analysis is the main component of a typical Document Image Analysis (DIA) system and plays an important role in pre-processing. However, regarding the Pashto language, the document images have not been explored so far. This research, for the first time, examines Pashto text along with graphics and proposes a deep learning-based classifier that can detect Pashto text and graphics per document. Another notable contribution of this research is the creation of a real dataset, which contains more than 1,000 images of the Pashto documents captured by a camera. For this dataset, we applied the convolution neural network (CNN) following a deep learning technique. Our intended method is based on the development of the advanced and classical variant of Faster R-CNN called Single-Shot Detector (SSD). The evaluation was performed by examining the 300 images from the test set. Through this way, we achieved a mean average precision (mAP) of 84.90%.
{"title":"Pashto script and graphics detection in camera captured Pashto document images using deep learning model","authors":"Khan Bahadar, Riaz Ahmad, Khursheed Aurangzeb, Siraj Muhammad, Khalil Ullah, Ibrar Hussain, Ikram Syed, Muhammad Shahid Anwar","doi":"10.7717/peerj-cs.2089","DOIUrl":"https://doi.org/10.7717/peerj-cs.2089","url":null,"abstract":"Layout analysis is the main component of a typical Document Image Analysis (DIA) system and plays an important role in pre-processing. However, regarding the Pashto language, the document images have not been explored so far. This research, for the first time, examines Pashto text along with graphics and proposes a deep learning-based classifier that can detect Pashto text and graphics per document. Another notable contribution of this research is the creation of a real dataset, which contains more than 1,000 images of the Pashto documents captured by a camera. For this dataset, we applied the convolution neural network (CNN) following a deep learning technique. Our intended method is based on the development of the advanced and classical variant of Faster R-CNN called Single-Shot Detector (SSD). The evaluation was performed by examining the 300 images from the test set. Through this way, we achieved a mean average precision (mAP) of 84.90%.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"12 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141773078","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Long reads generated by third-generation sequencing (3GS) technologies are involved in many biological analyses and play a vital role due to their ultra-long read length. However, the high error rate affects the downstream process. DeepCorr, a novel error correction algorithm for data from both PacBio and ONT platforms based on deep learning is proposed. The core algorithm adopts a recurrent neural network to capture the long-term dependencies in the long reads to convert the problem of long-read error correction to a multi-classification task. It first aligns the high-precision short reads to long reads to generate the corresponding feature vectors and labels, then feeds these vectors to the neural network, and finally trains the model for prediction and error correction. DeepCorr produces untrimmed corrected long reads and improves the alignment identity while maintaining the length advantage. It can capture and make full use of the dependencies to polish those bases that are not aligned by any short read. DeepCorr achieves better performance than that of the state-of-the-art error correction methods on real-world PacBio and ONT benchmark data sets and consumes fewer computing resources. It is a comprehensive deep learning-based tool that enables one to correct long reads accurately.
第三代测序（3GS）技术产生的长读数因其超长读数长度而在许多生物分析中发挥着重要作用。然而，高错误率会影响下游流程。DeepCorr 是一种基于深度学习的新型纠错算法，适用于 PacBio 和 ONT 平台的数据。其核心算法采用递归神经网络捕捉长读数中的长期依赖关系，将长读数纠错问题转化为多分类任务。它首先将高精度短读与长读对齐，生成相应的特征向量和标签，然后将这些向量输入神经网络，最后训练模型进行预测和纠错。DeepCorr 可以生成未经修剪的校正长读数，并在保持长度优势的同时提高比对识别率。它可以捕捉并充分利用依赖关系，打磨那些没有被任何短读数配准的碱基。在实际的 PacBio 和 ONT 基准数据集上，DeepCorr 比最先进的纠错方法取得了更好的性能，而且消耗的计算资源更少。它是一种基于深度学习的综合工具，能准确校正长读数。
{"title":"DeepCorr: a novel error correction method for 3GS long reads based on deep learning","authors":"Rongshu Wang, Jianhua Chen","doi":"10.7717/peerj-cs.2160","DOIUrl":"https://doi.org/10.7717/peerj-cs.2160","url":null,"abstract":"Long reads generated by third-generation sequencing (3GS) technologies are involved in many biological analyses and play a vital role due to their ultra-long read length. However, the high error rate affects the downstream process. DeepCorr, a novel error correction algorithm for data from both PacBio and ONT platforms based on deep learning is proposed. The core algorithm adopts a recurrent neural network to capture the long-term dependencies in the long reads to convert the problem of long-read error correction to a multi-classification task. It first aligns the high-precision short reads to long reads to generate the corresponding feature vectors and labels, then feeds these vectors to the neural network, and finally trains the model for prediction and error correction. DeepCorr produces untrimmed corrected long reads and improves the alignment identity while maintaining the length advantage. It can capture and make full use of the dependencies to polish those bases that are not aligned by any short read. DeepCorr achieves better performance than that of the state-of-the-art error correction methods on real-world PacBio and ONT benchmark data sets and consumes fewer computing resources. It is a comprehensive deep learning-based tool that enables one to correct long reads accurately.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"58 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141773079","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Every work environment contains different types of risks and interactions between risks. Therefore, the method to be used when making a risk assessment is very important. When determining which risk assessment method (RAM) to use, there are many factors such as the types of risks in the work environment, the interactions of these risks with each other, and their distance from the employees. Although there are many RAMs available, there is no RAM that will suit all workplaces and which method to choose is the biggest question. There is no internationally accepted scale or trend on this subject. In the study, 26 sectors, 10 different RAMs and 10 criteria were determined. A hybrid approach has been designed to determine the most suitable RAMs for sectors by using k-means clustering and support vector machine (SVM) classification algorithms, which are machine learning (ML) algorithms. First, the data set was divided into subsets with the k-means algorithm. Then, the SVM algorithm was run on all subsets with different characteristics. Finally, the results of all subsets were combined to obtain the result of the entire dataset. Thus, instead of the threshold value determined for a single and large cluster affecting the entire cluster and being made mandatory for all of them, a flexible structure was created by determining separate threshold values for each sub-cluster according to their characteristics. In this way, machine support was provided by selecting the most suitable RAMs for the sectors and eliminating the administrative and software problems in the selection phase from the manpower. The first comparison result of the proposed method was found to be the hybrid method: 96.63%, k-means: 90.63 and SVM: 94.68%. In the second comparison made with five different ML algorithms, the results of the artificial neural networks (ANN): 87.44%, naive bayes (NB): 91.29%, decision trees (DT): 89.25%, random forest (RF): 81.23% and k-nearest neighbours (KNN): 85.43% were found.
{"title":"A hybrid approach based on k-means and SVM algorithms in selection of appropriate risk assessment methods for sectors","authors":"Fatih Topaloglu","doi":"10.7717/peerj-cs.2198","DOIUrl":"https://doi.org/10.7717/peerj-cs.2198","url":null,"abstract":"Every work environment contains different types of risks and interactions between risks. Therefore, the method to be used when making a risk assessment is very important. When determining which risk assessment method (RAM) to use, there are many factors such as the types of risks in the work environment, the interactions of these risks with each other, and their distance from the employees. Although there are many RAMs available, there is no RAM that will suit all workplaces and which method to choose is the biggest question. There is no internationally accepted scale or trend on this subject. In the study, 26 sectors, 10 different RAMs and 10 criteria were determined. A hybrid approach has been designed to determine the most suitable RAMs for sectors by using k-means clustering and support vector machine (SVM) classification algorithms, which are machine learning (ML) algorithms. First, the data set was divided into subsets with the k-means algorithm. Then, the SVM algorithm was run on all subsets with different characteristics. Finally, the results of all subsets were combined to obtain the result of the entire dataset. Thus, instead of the threshold value determined for a single and large cluster affecting the entire cluster and being made mandatory for all of them, a flexible structure was created by determining separate threshold values for each sub-cluster according to their characteristics. In this way, machine support was provided by selecting the most suitable RAMs for the sectors and eliminating the administrative and software problems in the selection phase from the manpower. The first comparison result of the proposed method was found to be the hybrid method: 96.63%, k-means: 90.63 and SVM: 94.68%. In the second comparison made with five different ML algorithms, the results of the artificial neural networks (ANN): 87.44%, naive bayes (NB): 91.29%, decision trees (DT): 89.25%, random forest (RF): 81.23% and k-nearest neighbours (KNN): 85.43% were found.","PeriodicalId":54224,"journal":{"name":"PeerJ Computer Science","volume":"67 1","pages":""},"PeriodicalIF":3.8,"publicationDate":"2024-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141773084","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：