Optimal Classification Research Articles

Methodology of optimal classification of regions by the level of industry development

The article describes the methodology of classification of multidimensional observations, according to which, in theory, development indices of regional industrial production were developed. According to this index, conditional regions are divided into lower, middle and upper classes. In the study, the development tendencies of the regions classified by the development index into stratified groups were also determined. At the end of the article, the theoretical and empirical conclusions obtained as a result of the research are presented.

Dendritic Cell Algorithm with Grouping Genetic Algorithm for Input Signal Generation

The artificial immune system, an excellent prototype for developing Machine Learning, is inspired by the function of the powerful natural immune system. As one of the prevalent classifiers, the Dendritic Cell Algorithm (DCA) has been widely used to solve binary problems in the real world. The classification of DCA depends on a data pre-processing procedure to generate input signals, where feature selection and signal categorization are the main work. However, the results of these studies also show that the signal generation of DCA is relatively weak, and all of them utilized a filter strategy to remove unimportant attributes. Ignoring filtered features and applying expertise may not produce an optimal classification result. To overcome these limitations, this study models feature selection and signal categorization into feature grouping problems. This study hybridizes Grouping Genetic Algorithm (GGA) with DCA to propose a novel DCA version, GGA-DCA, for accomplishing feature selection and signal categorization in a search process. The GGA-DCA aims to search for the optimal feature grouping scheme without expertise automatically. In this study, the data coding and operators of GGA are redefined for grouping tasks. The experimental results show that the proposed algorithm has significant advantages over the compared DCA expansion algorithms in terms of signal generation.

Diagnosing Noise-Induced Hearing Loss Sustained During Military Service Using Deep Neural Networks.

The diagnosis of noise-induced hearing loss (NIHL) is based on three requirements: a history of exposure to noise with the potential to cause hearing loss; the absence of known causes of hearing loss other than noise exposure; and the presence of certain features in the audiogram. All current methods for diagnosing NIHL have involved examination of the typical features of the audiograms of noise-exposed individuals and the formulation of quantitative rules for the identification of those features. This article describes an alternative approach based on the use of multilayer perceptrons (MLPs). The approach was applied to databases containing the ages and audiograms of individuals claiming compensation for NIHL sustained during military service (M-NIHL), who were assumed mostly to have M-NIHL, and control databases with no known exposure to intense sounds. The MLPs were trained so as to classify individuals as belonging to the exposed or control group based on their audiograms and ages, thereby automatically identifying the features of the audiogram that provide optimal classification. Two databases (noise exposed and nonexposed) were used for training and validation of the MLPs and two independent databases were used for evaluation and further analyses. The best-performing MLP was one trained to identify whether or not an individual had M-NIHL based on age and the audiogram for both ears. This achieved a sensitivity of 0.986 and a specificity of 0.902, giving an overall accuracy markedly higher than for previous methods.

D2D Cooperative Communication Network Resource Allocation Algorithm Based on Improved Monte Carlo Tree Search

In recent years, with the rapid development of mobile communication, D2D (Device-to-Device, D2D) cooperative communication network has become the main component of future communication network, which greatly improves the spectrum efficiency of the network and the quality of user communication. However, the existing D2D network resource allocation schemes have some problems, such as weak dynamic resource allocation capability and low user communication quality. In view of this challenge, this paper proposes a resource allocation algorithm for D2D cooperative communication networks based on improved Monte Carlo tree search. First, a double-chain deep deciduous Monte Carlo tree search (Dcdd-MCTS) resource allocation model is established, Then, the loss function composed of deciduous MCTS and parallel convolution network is used to update the parameters of the deep neural network model of Dcdd-MCTS. Then, the theory of optimal classification is used to solve the user’s transmit power. Finally, the optimal scheme of dynamic output resource allocation is output. The simulation results show that Dcdd-MCTS has good convergence. In the research on the distance between devices, compared with single-chain deep MCTS and joint optimization algorithm [52], the proposed algorithm in this paper increases the system throughput by 5%, 2%, respectively, and reduces the outage probability by 33%,18%.

Identification of bacteria in mixed infection from urinary tract of patient's samples using Raman analysis of dried droplets.

Urinary tract infections (UTI) are among the most frequent nosocomial infections. A fast identification of the pathogen and assignment of Gram type could help to prescribe most suitable treatments. Raman spectroscopy holds high potential for fast and reliable bacterial pathogens identification. While most studies so far have focused on individual pathogens or artificial mixtures, this contribution aims to translate the analysis to primary urine samples from patients with suspected UTIs. For this, we have included 59 primary urine samples out of which 29 were diagnosed as mixed infections. For Raman analysis, we first trained two classification models based on principal component analysis - linear discriminant analysis (PCA-LDA) with more than 3500 Raman spectra of 85 clinical isolates from 23 species in order to (1) identify the Gram type of the bacteria and (2) assign family membership to one of the six most abundant bacterial families in urinary tract infections (Enterobacteriaceae, Morganellaceae, Pseudomonadaceae, Enterococcaceae, Staphylococcaceae and Streptococcaceae). The classification models were applied to artificial mixtures of Gram positive and Gram negative bacteria to correctly predict mixed infections with an accuracy of 75%. Raman scans of dried droplets did not yet yield optimal classification results on family level. When translating the method to primary urine samples, we observed a strong bias towards Gram negative bacteria, on family level towards Morganellaceae, which reduced prediction accuracy. Spectral differences were observed between isolates grown on standard growth medium and bacteria of the same strain when characterized directly from the patient. Thus, improvement of the classification accuracy is expected with a larger data base containing also bacteria measured directly from the urine sample.

Enhancing fabric quality control: Implementing real-time defect detection with image processing techniques and arduino

The textile industry is a rapidly growing sector globally and plays a momentous role in many sectors like manufacturing, employment, and business operations in many developed countries. Cloth flaws account for over 85% of the failures experienced in the garment industry. Efforts are currently underway to enhance fabric consistency, making the identification of defects a critical step in the textile manufacturing process. However, the traditional manual inspection technique for detecting cloth flaws is time-consuming and labor-consuming. Consequently, automation has been introduced through image processing technology. This approach utilizes image processing techniques in MATLAB to locate faults, with fault detection carried out using an Arduino. To improve the accuracy of fabric defect identification, an electronic fabric inspection method has been proposed. This framework incorporates image processing techniques, employing MATLAB, and real-time applications implemented on the Arduino kit. Neural Networks serve as the optimal classifiers for fault classification. Upon detecting a flaw in the fabric, the system breaks shortly to remove the defective component. The identified fault is then displayed on the LCD, and the buzzer is activated.

Improving Generalized Zero-Shot Learning SSVEP Classification Performance From Data-Efficient Perspective.

Generalized zero-shot learning (GZSL) has significantly reduced the training requirements for steady-state visual evoked potential (SSVEP) based brain-computer interfaces (BCIs). Traditional methods require complete class data sets for training, but GZSL allows for only partial class data sets, dividing them into 'seen' (those with training data) and 'unseen' classes (those without training data). However, inefficient utilization of SSVEP data limits the accuracy and information transfer rate (ITR) of existing GZSL methods. To this end, we proposed a framework for more effective utilization of SSVEP data at three systematically combined levels: data acquisition, feature extraction, and decision-making. First, prevalent SSVEP-based BCIs overlook the inter-subject variance in visual latency and employ fixed sampling starting time (SST). We introduced a dynamic sampling starting time (DSST) strategy at the data acquisition level. This strategy uses the classification results on the validation set to find the optimal sampling starting time (OSST) for each subject. In addition, we developed a Transformer structure to capture the global information of input data for compensating the small receptive field of existing networks. The global receptive fields of the Transformer can adequately process the information from longer input sequences. For the decision-making level, we designed a classifier selection strategy that can automatically select the optimal classifier for the seen and unseen classes, respectively. We also proposed a training procedure to make the above solutions in conjunction with each other. Our method was validated on three public datasets and outperformed the state-of-the-art (SOTA) methods. Crucially, we also outperformed the representative methods that require training data for all classes.

Neural Correlation of EEG and Eye Movement in Natural Grasping Intention Estimation.

Decoding the user's natural grasp intent enhances the application of wearable robots, improving the daily lives of individuals with disabilities. Electroencephalogram (EEG) and eye movements are two natural representations when users generate grasp intent in their minds, with current studies decoding human intent by fusing EEG and eye movement signals. However, the neural correlation between these two signals remains unclear. Thus, this paper aims to explore the consistency between EEG and eye movement in natural grasping intention estimation. Specifically, six grasp intent pairs are decoded by combining feature vectors and utilizing the optimal classifier. Extensive experimental results indicate that the coupling between the EEG and eye movements intent patterns remains intact when the user generates a natural grasp intent, and concurrently, the EEG pattern is consistent with the eye movements pattern across the task pairs. Moreover, the findings reveal a solid connection between EEG and eye movements even when taking into account cortical EEG (originating from the visual or motor cortex) and the presence of a suboptimal classifier. Overall, this work uncovers the coupling correlation between EEG and eye movements and provides a reference for intention estimation.

Feature Selection with Optimal Variational Auto Encoder for Financial Crisis Prediction

Financial crisis prediction (FCP) received significant attention in the financial sector for decision-making. Proper forecasting of the number of firms possible to fail is important to determine the growth index and strength of a nation’s economy. Conventionally, numerous approaches have been developed in the design of accurate FCP processes. At the same time, classifier efficacy and predictive accuracy are inadequate for real-time applications. In addition, several established techniques carry out well to any of the specific datasets but are not adjustable to distinct datasets. Thus, there is a necessity for developing an effectual prediction technique for optimum classifier performance and adjustable to various datasets. This paper presents a novel multi-vs. optimization (MVO) based feature selection (FS) with an optimal variational auto encoder (OVAE) model for FCP. The proposed multi-vs. optimization based feature selection with optimal variational auto encoder (MVOFS-OVAE) model mainly aims to accomplish forecasting the financial crisis. For achieving this, the proposed MVOFS-OVAE model primarily pre-processes the financial data using min-max normalization. In addition, the MVOFS-OVAE model designs a feature subset selection process using the MVOFS approach. Followed by, the variational auto encoder (VAE) model is applied for the categorization of financial data into financial crisis or non-financial crisis. Finally, the differential evolution (DE) algorithm is utilized for the parameter tuning of the VAE model. A series of simulations on the benchmark dataset reported the betterment of the MVOFS-OVAE approach over the recent state of art approaches.

Evaluating the Effectiveness of Interactive Video Learning by Examining Machine Learning Classifiers Models: Graduate Students’ Perspectives

Important elements had an impact on how traditional learning was implemented and motivated researchers to develop Interactive Video Learning Effectiveness (IVL-E). These variables range from price to learning-environment to learner perspective, among others. This paper’s major objectives are to: (i) assess the effectiveness of Interactive Video Learning (IVL-E) using classification techniques and considering graduate students’ viewpoints, (ii) establish appropriate classification parameters to choose the optimum classifier model, and (iii) review prior works pertaining to IVL-E assessment. The study dataset is a sample of 63 datapoints randomly chosen by a survey performed at the College of Education, University of Bisha, Bisha, Saudi Arabia. A total of 123 registered postgraduate students made up the study population when using Google’s online questionnaire method, after all the respondents voluntarily agreed to fill out and submit the questionnaire. This study develops a reliable machine learning classifier’s model for classifying IVL-E. The created models use a backpropagation algorithm and are a type of multilayer classification perceptron. The best classification output was “interactive video learning performance measure”, which provided the highest results under: 1) support vector machine-based classifier (SVC), 2) decision tree (DT), and 3) light gradient-boosting machine classifier (lgb.LGBMClassifier). Regarding classification measures like balanced accuracy (high BCCR = 0.875), balanced error rate (low BER = 0.125), and optimization precision (highest OP = 0.999), our models performed extremely well according to the literature review.

Prediction and risk assessment of sepsis-associated encephalopathy in ICU based on interpretable machine learning

Sepsis-associated encephalopathy (SAE) is a major complication of sepsis and is associated with high mortality and poor long-term prognosis. The purpose of this study is to develop interpretable machine learning models to predict the occurrence of SAE after ICU admission and implement the individual prediction and analysis. Patients with sepsis admitted to ICU were included. SAE was diagnosed as glasgow coma score (GCS) less than 15. Statistical analysis at baseline was performed between SAE and non-SAE. Six machine learning classifiers were employed to predict the occurrence of SAE, and the adjustment of model super parameters was performed by using Bayesian optimization method. Finally, the optimal algorithm was selected according to the prediction efficiency. In addition, professional physicians were invited to evaluate our model prediction results for further quantitative assessment of the model interpretability. The preliminary analysis of variance showed significant differences in the incidence of SAE among patients with pathogen infection. There were significant differences in physical indicators like respiratory rate, temperature, SpO2 and mean arterial pressure (P < 0.001). In addition, the laboratory results were also significantly different. The optimal classification model (XGBoost) indicated that the best risk factors (cut-off points) were creatinine (1.1 mg/dl), mean respiratory rate (18), pH (7.38), age (72), chlorine (101 mmol/L), sodium (138.5 k/ul), SAPSII score (23), platelet count (160), and phosphorus (2.4 and 5.0 mg/dL). The ranked features derived from the best model (AUC is 0.8837) were mechanical ventilation, duration of mechanical ventilation, phosphorus, SOFA score, and vasopressin usage. The SAE risk prediction model based on XGBoost created here can make very accurate predictions using simple indicators and support the visual explanation. The interpretable model was effectively evaluated by professional physicians and can help them predict the occurrence of SAE more intuitively.

A Deep Learning Model for Preoperative Differentiation of Glioblastoma, Brain Metastasis, and Primary Central Nervous System Lymphoma: An External Validation Study

(1) Background: Neuroimaging differentiation of glioblastoma, primary central nervous system lymphoma (PCNSL) and solitary brain metastasis (BM) represents a diagnostic and therapeutic challenge in neurosurgical practice, expanding the burden of care and exposing patients to additional risks related to further invasive procedures and treatment delays. In addition, atypical cases and overlapping features have not been entirely addressed by modern diagnostic research. The aim of this study was to validate a previously designed and internally validated ResNet101 deep learning model to differentiate glioblastomas, PCNSLs and BMs. (2) Methods: We enrolled 126 patients (glioblastoma: n = 64; PCNSL: n = 27; BM: n = 35) with preoperative T1Gd-MRI scans and histopathological confirmation. Each lesion was segmented, and all regions of interest were exported in a DICOM dataset. A pre-trained ResNet101 deep neural network model implemented in a previous work on 121 patients was externally validated on the current cohort to differentiate glioblastomas, PCNSLs and BMs on T1Gd-MRI scans. (3) Results: The model achieved optimal classification performance in distinguishing PCNSLs (AUC: 0.73; 95%CI: 0.62–0.85), glioblastomas (AUC: 0.78; 95%CI: 0.71–0.87) and moderate to low ability in differentiating BMs (AUC: 0.63; 95%CI: 0.52–0.76). The performance of expert neuro-radiologists on conventional plus advanced MR imaging, assessed by retrospectively reviewing the diagnostic reports of the selected cohort of patients, was found superior in accuracy for BMs (89.69%) and not inferior for PCNSL (82.90%) and glioblastomas (84.09%). (4) Conclusions: We investigated whether the previously published deep learning model was generalizable to an external population recruited at a different institution—this validation confirmed the consistency of the model and laid the groundwork for future clinical applications in brain tumour classification. This artificial intelligence-based model might represent a valuable educational resource and, if largely replicated on prospective data, help physicians differentiate glioblastomas, PCNSL and solitary BMs, especially in settings with limited resources.

Improved Mask R-CNN Segmentation and Bayesian Interactive Adaboost CNN Classification for Breast Cancer Detection on Bach Dataset

Breast cancer is considered as the predominant type of cancer that affects more than ten percentage of the worldwide female population. Though microscopic evaluation remains to be a significant method for diagnosing, time and cost complexity seeks alternative and effective computer aided design for rapid and more accurate detection of the disease. As DL (Deep Learning) possess a significant contribution in accomplishing machine automation, this study intends to resolve existing problems with regard to lack of accuracy by proposing DL based algorithms. The study proposes Improved-Mask R CNN (I-MRCNN) method for segmentation. In this process, RPN (Region Proposal Network), predicts the objectless scores and object bound at every position. Here, (RoI Align) Region of interest Align is used for feature extraction as it is capable of resolving the conventional RoI pooling issues by attaining high accuracy for small objects and also eliminates quantization issues. Further, classification is performed using the proposed Bayesian Interactive Adaboost CNN classifier (B-IAB- CNN) that integrates the advantages of CNN, Bayesian and Adaboost classifier. The advantages of the three classifier enable optimum classification of the input Bach dataset that is confirmed through the results of performance analysis of the proposed system. Outcomes reveal that, average accuracy for segmentation is 96.32%, while, the classification accuracy is exposed to be 96%. As Timely prediction is significant, high prediction rate of the proposed system will assist the medical practitioners to detect breast cancer quickly which is the important practical implication from this study for diagnosing breast cancer.

SOLVING THE PROBLEM OF DETECTING PHISHING WEBSITES USING ENSEMBLE LEARNING MODELS

Due to the popularity of the easiest way to obtain personal information among attackers, phishing detection is becoming a popular area for research aimed at countering the implementation of such attacks. Malicious website detection is essential to prevent the spread of malware and protect end users from victims. Unfortunately, malicious URL detection still needs to be better understood due to a lack of features and inaccurate classification. Possible sources were examined in order to investigate the subject. Based on the collected information from previous studies, this study is devoted to solving the problem of detecting phishing websites using Ensemble Learning. The aim of the work is to choose the most optimal algorithm for classifying phishing websites using gradient boosting algorithms. AdaBoost, CatBoost, and Gradient Boosting Classifier were chosen as Ensemble Learning algorithms and were used to improve the efficiency of classifiers. Practical studies of the parameters of each algorithm for finding the optimal classification model are given. Research and experiments were carried out on a dataset containing information extracted from the contents of a URL: main URL, domain, directory, and file. A thorough Exploratory Data Analysis (EDA) was carried out, as a result of which the main dependencies and patterns of determining phishing resources were identified using correlation analysis. ROC AUC Score was chosen as an evaluation metric for the algorithms. The best result for predicting phishing websites was demonstrated by the AdaBoost Classifier algorithm, with an average ROC AUC score of 99%. The results of the experiments were illustrated in the form of graphs and tables.

An Optimal Approach for Heart Sound Classification Using Grid Search in Hyperparameter Optimization of Machine Learning.

Heart-sound auscultation is one of the most widely used approaches for detecting cardiovascular disorders. Diagnosing abnormalities of heart sound using a stethoscope depends on the physician's skill and judgment. Several studies have shown promising results in automatically detecting cardiovascular disorders based on heart-sound signals. However, the accuracy performance needs to be enhanced as automated heart-sound classification aids in the early detection and prevention of the dangerous effects of cardiovascular problems. In this study, an optimal heart-sound classification method based on machine learning technologies for cardiovascular disease prediction is performed. It consists of three steps: pre-processing that sets the 5 s duration of the PhysioNet Challenge 2016 and 2022 datasets, feature extraction using Mel frequency cepstrum coefficients (MFCC), and classification using grid search for hyperparameter tuning of several classifier algorithms including k-nearest neighbor (K-NN), random forest (RF), artificial neural network (ANN), and support vector machine (SVM). The five-fold cross-validation was used to evaluate the performance of the proposed method. The best model obtained classification accuracy of 95.78% and 76.31%, which was assessed using PhysioNet Challenge 2016 and 2022, respectively. The findings demonstrate that the suggested approach obtained excellent classification results using PhysioNet Challenge 2016 and showed promising results using PhysioNet Challenge 2022. Therefore, the proposed method has been potentially developed as an additional tool to facilitate the medical practitioner in diagnosing the abnormality of the heart sound.

Optimization Analysis Model Determining PNMP Mandiri Loan Status Based on Pearson Correlation

PNPM Mandiri is an organization engaged in financing small and medium enterprises in the community. The problem that always occurs is an error in determining the loan status resulting in bad credit. This study aims to present a classification analysis model for determining loan status at PNPM Mandiri. The classification analysis model was built using the Perceptron algorithm artificial neural network. The analysis model will later be optimized using the Person Correlation (PC) method to measure the accuracy of the variables used. The research dataset is based on historical data from the last 2 years as many as 67 data samples. The analysis variables consist of Business Type (X1), Loan Amount (X2), Collateral (X3), Income (X4), and Expenses (X5). The results of the analysis show that the model built can provide optimal classification results. These results can be seen based on the results of variable measurements using the PC method indicating that variable X2 has no significant relationship. With the results of these measurements, the performance of the artificial neural network presents maximum results in determining loan status. Overall, the results of this study can provide an effective analytical model as well as an alternative solution for determining loan status.&#x0D;

On the optimal binary classifier with an application

The alternative accumulated improvement curve stochastic order is a criterion for the comparison of the performance of classifiers that predict binary responses. An explicit optimal classifier for this criterion is obtained. That optimal classifier has the largest ROC and CAP curves and indexes, that is, it is also optimal for the criteria based on the comparison of such curves and indexes. An application of the results to the search of the best classifier to predict clients of a bank which will make a transaction in the future is developed.

Diffusion-weighted imaging and arterial spin labeling radiomics features may improve differentiation between radiation-induced brain injury and glioma recurrence.

To determine whether radiomics features derived from diffusion-weighted imaging (DWI) and arterial spin labeling (ASL) can improve the differentiation between radiation-induced brain injury (RIBI) and tumor recurrence (TR) in glioma patients. A total of 4199 radiomics features were extracted from conventional MRI, apparent diffusion coefficient (ADC), and cerebral blood flow (CBF) maps, obtained from 96 pathologically confirmed WHO grade 2~4 gliomas with enhancement after standard treatment. The intraclass correlation coefficient (ICC) was used to test segmentation stability between two doctors. Radiomics features were selected using the Mann-Whitney U test, LASSO regression, and RFE algorithms. Four machine learning classifiers were adopted to establish radiomics models. The diagnostic performance of multiparameter, conventional, and single-parameter MRI radiomics models was compared using the area under the curve (AUC). The models were evaluated in the subsequent independent validation set (n = 30). Eight important radiomics features (3 from conventional MRI, 1 from ADC, and 4 from CBF) were selected. Support vector machine (SVM) was chosen as the optimal classifier. The diagnostic performance of the multiparameter MRI radiomics model (AUC 0.96) was higher than that of the conventional MRI (AUC 0.88), ADC (AUC 0.91), and CBF (AUC 0.95) radiomics models. For subgroup analysis, the multiparameter MRI radiomics model showed similar performance, with AUCs of 0.98 in WHO grade 2~3 and 0.96 in WHO grade 4. The incorporation of noninvasive DWI and ASL into the MRI radiomics model improved the diagnostic performance in differentiating RIBI from TR; ASL, especially, played a significant role. • The multiparameter MRI radiomics model was superior to the conventional MRI radiomics model in differentiating glioma recurrence from radiation-induced brain injury. • Diffusion and perfusion MRI could improve the ability of the radiomics model in predicting the progression in patients with glioma. • Arterial spin labeling played an important role in predicting glioma progression using radiomics models.

Early detection of liver disorders using hybrid soft computing techniques for optimal feature selection and classification

SummaryHuman liver disease is a sort of illness that starts in the ovaries and is particularly dangerous for women. As a consequence, aberrant cells develop that have the potential to spread to other parts of the body. Liver disease is a serious disorder that affects women's ovaries and is difficult to identify early on, which is why it is still one of the leading causes of mortality. The significance of unequivocal confirmation of intrinsic and typical components in establishing new frameworks to detect and eliminate danger is substantial. In this study, we present a hybrid soft computing approach for early diagnosis of liver problems (EDLD‐HS). For high and low level feature extraction models, we first employ an improved ant swarm optimization (IASO) approach. Then, for optimum feature selection, we develop a modified whale search optimization (MWSO) technique that combines features that may reflect both texture patterns and semantic backdrop scattered in the data. After that, we employed a hybrid swallow swarm intelligent‐deep neural network (HSSI‐DNN) classifier to determine the stage of liver illness. Finally, we used MATLAB R2014a to test our proposed EDLD‐HS method using well‐known benchmark datasets including BUPA, ILPD, and MPRLPD. The simulation findings are compared to existing state‐of‐the‐art methodologies in terms of accuracy, specificity, sensitivity, precision, recall, F1‐score, G‐mean, and area under curve (AUC). The detection accuracy of the proposed HSSI‐DNN classifier is 83.26% (BUPA), 84.26% (ILPD), and 91.23% (ILPD) (MPRLPD).

Crop type mapping in the central part of the North China Plain using Sentinel-2 time series and machine learning

The North China Plain (NCP), a major agricultural area in China, plays an important role in China's grain production. Timely and accurate crop information for NCP is very important to China's food security and sustainable development. Due to high variability of the temporal profiles of vegetation indices, classification models using temporally aggregated remote sensing data often exhibit suboptimal performance for multi-crop classification in the NCP with complex cropping patterns. Therefore, optimal feature sets and classification models should be developed for efficient and accurate crop mapping in this region. In this study, we used all available Sentinel-2 imagery in 2020 to map major crops including winter wheat/corn, cotton, peanut, and millet in a typical winter wheat production city in the central part of the NCP. NDVI time series, textural, and phenological features from Sentinel-2 time series and topographic features of the study area were derived as input features (394 features in total). Two feature selection methods, random forest and unsupervised feature selection based on multi-subspace randomization and collaboration (SRCFS), were used to select 20 informative features from the 394 features. Then, four groups of features were evaluated with three machine learning classifiers, i.e., random forest (RF), support vector machine (SVM), and artificial neural network (ANN). The results showed that the most useful features for crop type classification for the region were phenological and textural features during February to March and April to May. When using the full feature set, RF provided the best results compared with SVM and ANN. However, both RF and SVM classifier with 20 RF-selected features generated the optimal results. The crops were identified with an overall accuracy of 93% and a kappa coefficient of 0.9 in the final 10-meter resolution crop map. The feature selection and machine learning classification methods can be applied to high-resolution crop mapping using time series of Sentinel-2 data in agricultural regions with mixed cropping patterns in an efficient manner.