Annealing strategy for an enhance rule pruning technique in ACO-Based rule classification

Ant colony optimization (ACO) was successfully applied to data mining classification task through ant-mining algorithms. Exploration and exploitation are search strategies that guide the learning process of a classification model and generate a list of rules. Exploitation refers to the process of intensifying the search for neighbors in good regions, whereas exploration aims towards new promising regions during a search process. The existing balance between exploration and exploitation in the rule construction procedure is limited to the roulette wheel selection mechanism, which complicates rule generation. Thus, low-coverage complex rules with irrelevant terms will be generated. This work proposes an enhancement rule pruning procedure for the ACO algorithm that can be used in rule-based classification. This procedure, called the annealing strategy, is an improvement of ant-mining algorithms in the rule construction procedure. Presented as a pre-pruning technique, the annealing strategy deals first with irrelevant terms before creating a complete rule through an annealing schedule. The proposed improvement was tested through benchmarking experiments, and results were compared with those of four of the most related ant-mining algorithms, namely, Ant-Miner, CAnt-Miner, TACO-Miner, and Ant-Miner with hybrid pruner. Results display that our proposed technique achieves better performance in terms of classification accuracy, model size, and computational time. The proposed annealing schedule can be used in other ACO variants for different applications to improve classification accuracy.

Download Full-text

Tobacco Leaf Grading Based on Deep Convolutional Neural Networks and Machine Vision

Journal of the ASABE ◽

10.13031/ja.14537 ◽

2021 ◽

Vol 65 (1) ◽

pp. 11-22

Author(s):

Mengyao Lu ◽

Shuwen Jiang ◽

Cong Wang ◽

Dong Chen ◽

Tian’en Chen

Keyword(s):

Neural Networks ◽

Deep Learning ◽

Transfer Learning ◽

Convolutional Neural Networks ◽

Classification Accuracy ◽

Classification Model ◽

List Type ◽

Tobacco Leaves ◽

Tobacco Leaf ◽

Grading Model

HighlightsA classification model for the front and back sides of tobacco leaves was developed for application in industry.A tobacco leaf grading method that combines a CNN with double-branch integration was proposed.The A-ResNet network was proposed and compared with other classic CNN networks.The grading accuracy of eight different grades was 91.30% and the testing time was 82.180 ms, showing a relatively high classification accuracy and efficiency.Abstract. Flue-cured tobacco leaf grading is a key step in the production and processing of Chinese-style cigarette raw materials, directly affecting cigarette blend and quality stability. At present, manual grading of tobacco leaves is dominant in China, resulting in unsatisfactory grading quality and consuming considerable material and financial resources. In this study, for fast, accurate, and non-destructive tobacco leaf grading, 2,791 flue-cured tobacco leaves of eight different grades in south Anhui Province, China, were chosen as the study sample, and a tobacco leaf grading method that combines convolutional neural networks and double-branch integration was proposed. First, a classification model for the front and back sides of tobacco leaves was trained by transfer learning. Second, two processing methods (equal-scaled resizing and cropping) were used to obtain global images and local patches from the front sides of tobacco leaves. A global image-based tobacco leaf grading model was then developed using the proposed A-ResNet-65 network, and a local patch-based tobacco leaf grading model was developed using the ResNet-34 network. These two networks were compared with classic deep learning networks, such as VGGNet, GoogLeNet-V3, and ResNet. Finally, the grading results of the two grading models were integrated to realize tobacco leaf grading. The tobacco leaf classification accuracy of the final model, for eight different grades, was 91.30%, and grading of a single tobacco leaf required 82.180 ms. The proposed method achieved a relatively high grading accuracy and efficiency. It provides a method for industrial implementation of the tobacco leaf grading and offers a new approach for the quality grading of other agricultural products. Keywords: Convolutional neural network, Deep learning, Image classification, Transfer learning, Tobacco leaf grading

Download Full-text

RESEARCH ON BRAIN NETWORK CLASSIFICATION METHOD BASED ON INTEGRATED MODEL

10.36106/paripex/9308196 ◽

2020 ◽

pp. 1-2

Author(s):

Zhang- sensen

Keyword(s):

Machine Learning ◽

Classification Accuracy ◽

Brain Network ◽

Classification Model ◽

Comparative Experiment ◽

Machine Learning Methods ◽

Healthy Elderly ◽

Healthy Elderly People ◽

The Brain

mild cognitive impairment (MCI) is a condition between healthy elderly people and alzheimer's disease (AD). At present, brain network analysis based on machine learning methods can help diagnose MCI. In this paper, the brain network is divided into several subnets based on the shortest path,and the feature vectors of each subnet are extracted and classified. In order to make full use of subnet information, this paper adopts integrated classification model for classification.Each base classification model can predict the classification of a subnet,and the classification results of all subnets are calculated as the classification results of brain network.In order to verify the effectiveness of this method,a brain network of 66 people was constructed and a comparative experiment was carried out.The experimental results show that the classification accuracy of the integrated classification model proposed in this paper is 19% higher than that of SVM,which effectively improves the classification accuracy

Download Full-text

Ship Classification in High-Resolution SAR Images Using Deep Learning of Small Datasets

Sensors ◽

10.3390/s18092929 ◽

2018 ◽

Vol 18 (9) ◽

pp. 2929 ◽

Cited By ~ 14

Author(s):

Yuanyuan Wang ◽

Chao Wang ◽

Hong Zhang

Keyword(s):

Deep Learning ◽

High Resolution ◽

Classification Accuracy ◽

Experimental Results ◽

Fine Tuning ◽

Classification Model ◽

Great Success ◽

Sar Images ◽

Convolutional Networks ◽

Ship Classification

With the capability to automatically learn discriminative features, deep learning has experienced great success in natural images but has rarely been explored for ship classification in high-resolution SAR images due to the training bottleneck caused by the small datasets. In this paper, convolutional neural networks (CNNs) are applied to ship classification by using SAR images with the small datasets. First, ship chips are constructed from high-resolution SAR images and split into training and validation datasets. Second, a ship classification model is constructed based on very deep convolutional networks (VGG). Then, VGG is pretrained via ImageNet, and fine tuning is utilized to train our model. Six scenes of COSMO-SkyMed images are used to evaluate our proposed model with regard to the classification accuracy. The experimental results reveal that (1) our proposed ship classification model trained by fine tuning achieves more than 95% average classification accuracy, even with 5-cross validation; (2) compared with other models, the ship classification model based on VGG16 achieves at least 2% higher accuracies for classification. These experimental results reveal the effectiveness of our proposed method.

Download Full-text

Overview of Predictive Modeling Approaches in Health Care Data Mining

Business Intelligence ◽

10.4018/978-1-4666-9562-7.ch004 ◽

2016 ◽

pp. 73-95 ◽

Cited By ~ 1

Author(s):

Sunita Soni

Keyword(s):

Data Mining ◽

Predictive Modeling ◽

Clinical Information ◽

Ease Of Use ◽

Classification Model ◽

Computational Time ◽

Data Sets ◽

Advantages And Disadvantages ◽

Mining Methods ◽

Predictive Data Mining

Medical data mining has great potential for exploring the hidden pattern in the data sets of the medical domain. A predictive modeling approach of Data Mining has been systematically applied for the prognosis, diagnosis, and planning for treatment of chronic disease. For example, a classification system can assist the physician to predict if the patient is likely to have a certain disease, or by considering the output of the classification model, the physician can make a better decision on the treatment to be applied to the patient. Once the model is evaluated and verified, it may be embedded within clinical information systems. The objective of this chapter is to extensively study the various predictive data mining methods to evaluate their usage in terms of accuracy, computational time, comprehensibility of the results, ease of use of the algorithm, and advantages and disadvantages to relatively naive medical users. The research has shown that there is not a single best prediction tool, but instead, the best performing algorithm will depend on the features of the dataset to be analyzed.

Download Full-text

A New Hybrid Feature Subset Selection Framework Based on Binary Genetic Algorithm and Information Theory

International Journal of Computational Intelligence and Applications ◽

10.1142/s1469026819500202 ◽

2019 ◽

Vol 18 (03) ◽

pp. 1950020 ◽

Cited By ~ 13

Author(s):

Alok Kumar Shukla ◽

Pradeep Singh ◽

Manu Vardhan

Keyword(s):

Genetic Algorithm ◽

Feature Selection ◽

Classification Accuracy ◽

B Cell Lymphoma ◽

Feature Subset Selection ◽

Classification Model ◽

Significant Feature ◽

Support Vector ◽

Feature Subset ◽

Binary Genetic Algorithm

The explosion of the high-dimensional dataset in the scientific repository has been encouraging interdisciplinary research on data mining, pattern recognition and bioinformatics. The fundamental problem of the individual Feature Selection (FS) method is extracting informative features for classification model and to seek for the malignant disease at low computational cost. In addition, existing FS approaches overlook the fact that for a given cardinality, there can be several subsets with similar information. This paper introduces a novel hybrid FS algorithm, called Filter-Wrapper Feature Selection (FWFS) for a classification problem and also addresses the limitations of existing methods. In the proposed model, the front-end filter ranking method as Conditional Mutual Information Maximization (CMIM) selects the high ranked feature subset while the succeeding method as Binary Genetic Algorithm (BGA) accelerates the search in identifying the significant feature subsets. One of the merits of the proposed method is that, unlike an exhaustive method, it speeds up the FS procedure without lancing of classification accuracy on reduced dataset when a learning model is applied to the selected subsets of features. The efficacy of the proposed (FWFS) method is examined by Naive Bayes (NB) classifier which works as a fitness function. The effectiveness of the selected feature subset is evaluated using numerous classifiers on five biological datasets and five UCI datasets of a varied dimensionality and number of instances. The experimental results emphasize that the proposed method provides additional support to the significant reduction of the features and outperforms the existing methods. For microarray data-sets, we found the lowest classification accuracy is 61.24% on SRBCT dataset and highest accuracy is 99.32% on Diffuse large B-cell lymphoma (DLBCL). In UCI datasets, the lowest classification accuracy is 40.04% on the Lymphography using k-nearest neighbor (k-NN) and highest classification accuracy is 99.05% on the ionosphere using support vector machine (SVM).

Download Full-text

A Novel Hybrid Approach for Chronic Disease Classification

International Journal of Healthcare Information Systems and Informatics ◽

10.4018/ijhisi.2020010101 ◽

2020 ◽

Vol 15 (1) ◽

pp. 1-19 ◽

Cited By ~ 1

Author(s):

Divya Jain ◽

Vijendra Singh

Keyword(s):

Chronic Disease ◽

Classification Accuracy ◽

Hybrid Approach ◽

Substantial Reduction ◽

Disease Classification ◽

Computational Time ◽

Second Phase ◽

Two Phase ◽

Hybrid Classification ◽

Diagnostic Framework

A two-phase diagnostic framework based on hybrid classification for the diagnosis of chronic disease is proposed. In the first phase, feature selection via ReliefF method and feature extraction via PCA method are incorporated. In the second phase, efficient optimization of SVM parameters via grid search method is performed. The proposed hybrid classification approach is then tested with seven popular chronic disease datasets using a cross-validation method. Experiments are then conducted to evaluate the presented classification method vis-à-vis four other existing classifiers that are applied on the same chronic disease datasets. Results show that the presented approach reduces approximately 40% of the extraneous and surplus features with substantial reduction in the execution time for mining all datasets, achieving the highest classification accuracy of 98.5%. It is concluded that with the presented approach, excellent classification accuracy is achieved for each chronic disease dataset while irrelevant and redundant features may be eliminated, thereby substantially reducing the diagnostic complexity and resulting computational time.

Download Full-text

An EEG-Based Identity Authentication System with Audiovisual Paradigm in IoT

Sensors ◽

10.3390/s19071664 ◽

2019 ◽

Vol 19 (7) ◽

pp. 1664 ◽

Cited By ~ 1

Author(s):

Haiping Huang ◽

Linkang Hu ◽

Fu Xiao ◽

Anming Du ◽

Ning Ye ◽

...

Keyword(s):

Classification Accuracy ◽

Reference Electrode ◽

Identity Authentication ◽

Classification Model ◽

Experimental Result ◽

Ensemble Averaging ◽

Security Risks ◽

Adaptive Feature Selection ◽

Authentication System ◽

Audiovisual Presentation

With the continuous increment of security risks and the limitations of traditional modes, it is necessary to design a universal and trustworthy identity authentication system for intelligent Internet of Things (IoT) applications such as an intelligent entrance guard. The characteristics of EEG (electroencephalography) have gained the confidence of researchers due to its uniqueness, stability, and universality. However, the limited usability of the experimental paradigm and the unsatisfactory classification accuracy have so far prevented the identity authentication system based on EEG to become commonplace in IoT scenarios. To address these problems, an audiovisual presentation paradigm is proposed to record the EEG signals of subjects. In the pre-processing stage, the reference electrode, ensemble averaging, and independent component analysis methods are used to remove artifacts. In the feature extraction stage, adaptive feature selection and bagging ensemble learning algorithms establish the optimal classification model. The experimental result shows that our proposal achieves the best classification accuracy when compared with other paradigms and typical EEG-based authentication methods, and the test evaluation on a login scenario is designed to further demonstrate that the proposed system is feasible, effective, and reliable.

Download Full-text

The CSP-Based New Features Plus Non-Convex Log Sparse Feature Selection for Motor Imagery EEG Classification

Sensors ◽

10.3390/s20174749 ◽

2020 ◽

Vol 20 (17) ◽

pp. 4749

Author(s):

Shaorong Zhang ◽

Zhibin Zhu ◽

Benxin Zhang ◽

Bao Feng ◽

Tianyou Yu ◽

...

Keyword(s):

Feature Extraction ◽

Feature Selection ◽

Motor Imagery ◽

Classification Accuracy ◽

Feature Selection Method ◽

Extraction Methods ◽

Extraction Time ◽

Classification Model ◽

Discrete Wavelet ◽

New Feature

The common spatial pattern (CSP) is a very effective feature extraction method in motor imagery based brain computer interface (BCI), but its performance depends on the selection of the optimal frequency band. Although a lot of research works have been proposed to improve CSP, most of these works have the problems of large computation costs and long feature extraction time. To this end, three new feature extraction methods based on CSP and a new feature selection method based on non-convex log regularization are proposed in this paper. Firstly, EEG signals are spatially filtered by CSP, and then three new feature extraction methods are proposed. We called them CSP-wavelet, CSP-WPD and CSP-FB, respectively. For CSP-Wavelet and CSP-WPD, the discrete wavelet transform (DWT) or wavelet packet decomposition (WPD) is used to decompose the spatially filtered signals, and then the energy and standard deviation of the wavelet coefficients are extracted as features. For CSP-FB, the spatially filtered signals are filtered into multiple bands by a filter bank (FB), and then the logarithm of variances of each band are extracted as features. Secondly, a sparse optimization method regularized with a non-convex log function is proposed for the feature selection, which we called LOG, and an optimization algorithm for LOG is given. Finally, ensemble learning is used for secondary feature selection and classification model construction. Combing feature extraction and feature selection methods, a total of three new EEG decoding methods are obtained, namely CSP-Wavelet+LOG, CSP-WPD+LOG, and CSP-FB+LOG. Four public motor imagery datasets are used to verify the performance of the proposed methods. Compared to existing methods, the proposed methods achieved the highest average classification accuracy of 88.86, 83.40, 81.53, and 80.83 in datasets 1–4, respectively. The feature extraction time of CSP-FB is the shortest. The experimental results show that the proposed methods can effectively improve the classification accuracy and reduce the feature extraction time. With comprehensive consideration of classification accuracy and feature extraction time, CSP-FB+LOG has the best performance and can be used for the real-time BCI system.

Download Full-text

Image-Based Artificial Intelligence Methods for Product Control of Tablet Coating Quality

Pharmaceutics ◽

10.3390/pharmaceutics12090877 ◽

2020 ◽

Vol 12 (9) ◽

pp. 877

Author(s):

Cosima Hirschberg ◽

Magnus Edinger ◽

Else Holmfred ◽

Jukka Rantanen ◽

Johan Boetker

Keyword(s):

Classification Model ◽

Computational Time ◽

Support Vector ◽

Coating Quality ◽

Coating Solution ◽

Human Decision ◽

Control Situations ◽

The Individual ◽

Product Control

Mimicking the human decision-making process is challenging. Especially, many process control situations during the manufacturing of pharmaceuticals are based on visual observations and related experience-based actions. The aim of the present work was to investigate the use of image analysis to classify the quality of coated tablets. Tablets with an increasing amount of coating solution were imaged by fast scanning using a conventional office scanner. A segmentation routine was implemented to the images, allowing the extraction of numeric image-based information from individual tablets. The image preprocessing was performed prior to utilization of four different classification techniques for the individual tablet images. The support vector machine (SVM) technique performed superior compared to a convolutional neural network (CNN) in relation to computational time, and this approach was also slightly better at classifying the tablets correctly. The fastest multivariate method was partial least squares (PLS) regression, but this method was hampered by the inferior classification accuracy of the tablets. Finally, it was possible to create a numerical threshold classification model with an accuracy comparable to the SVM approach, so it is evident that there exist multiple valid options for classifying coated tablets.

Download Full-text

A Novel Autonomous Perceptron Model for Pattern Classification Applications

Entropy ◽

10.3390/e21080763 ◽

2019 ◽

Vol 21 (8) ◽

pp. 763 ◽

Cited By ~ 12

Author(s):

Alaa Sagheer ◽

Mohammed Zidan ◽

Mohammed M. Abdelsamea

Keyword(s):

Pattern Classification ◽

Data Science ◽

Science Research ◽

Efficient Solutions ◽

Classification Model ◽

Computational Time ◽

Quantum Bit ◽

Training Samples ◽

Artificial Neural Network Ann ◽

Definition Of

Pattern classification represents a challenging problem in machine learning and data science research domains, especially when there is a limited availability of training samples. In recent years, artificial neural network (ANN) algorithms have demonstrated astonishing performance when compared to traditional generative and discriminative classification algorithms. However, due to the complexity of classical ANN architectures, ANNs are sometimes incapable of providing efficient solutions when addressing complex distribution problems. Motivated by the mathematical definition of a quantum bit (qubit), we propose a novel autonomous perceptron model (APM) that can solve the problem of the architecture complexity of traditional ANNs. APM is a nonlinear classification model that has a simple and fixed architecture inspired by the computational superposition power of the qubit. The proposed perceptron is able to construct the activation operators autonomously after a limited number of iterations. Several experiments using various datasets are conducted, where all the empirical results show the superiority of the proposed model as a classifier in terms of accuracy and computational time when it is compared with baseline classification models.

Download Full-text