Fig 2 - uploaded by Akizur Rahman
Content may be subject to copyright.
Source publication
In the fast-growing field of medicine and its dynamic demand in research, a study that proves significant improvement to healthcare seems imperative especially when it is on cancer research. This research paved way to such significant findings by the inclusion of feature selection as one of its major components. Feature selection has become a vital...
Contexts in source publication
Context 1
... study, the hidden layer used 18 neurons and 2 neurons in the output layer. ANNs used multiple layers for the training with an efficient method applied and with back propagation-learning algorithm to the network. Those outputs are probably standard or cancer. A feed-forward neural network necessary architecture representation presented below in Fig. 2. The feed-forward network has links extending to a single direction only. There is no backward connection in the feedforward network. All connections proceed from the input node toward the output node. The classifier needs to use for classification. This study used ANNs classifier for colon cancer classification. A typical Artificial ...
Context 2
... study, the hidden layer used 18 neurons and 2 neurons in the output layer. ANNs used multiple layers for the training with an efficient method applied and with back propagation-learning algorithm to the network. Those outputs are probably standard or cancer. A feed-forward neural network necessary architecture representation presented below in Fig. 2. The feed-forward network has links extending to a single direction only. There is no backward connection in the feed- forward network. All connections proceed from the input node toward the output node. The classifier needs to use for classification. This study used ANNs classifier for colon cancer classification. A typical Artificial ...
Similar publications
Not all features or attributes can be very helpful in image classification for all data. Inclusion of features that are irrelevant and redundant in the classification process may cripple the predictive accuracy of the model and may slow down the processing time. This paper aims to show that feature selection improves the training time, application...
Citations
... In addition, the authors have found that using the feature selection method may increase the classification accuracy of the model. However, the sample size was small [46]. ...
... • Machine Learning (ML): [5,20,[38][39][40]42,[44][45][46]50,51,53,54,57,59,62,63,65,68,69,71,72]. Table 1 shows the details of prediction models using ML algorithms. ...
... Text Dataset: [5,[38][39][40]42,46,50,51,57,62,68,72]. The authors used text datasets and they are summarized in Table 3. • Image Dataset: [9,10,[19][20][21][22]41,[43][44][45][47][48][49][52][53][54][55][56][58][59][60][61][63][64][65][66][67][69][70][71]. ...
One of the three most serious and deadly cancers in the world is colorectal cancer. The most crucial stage, like with any cancer, is early diagnosis. In the medical industry, artificial intelligence (AI) has recently made tremendous strides and showing promise for clinical applications. Machine learning (ML) and deep learning (DL) applications have recently gained popularity in the analysis of medical texts and images due to the benefits and achievements they have made in the early diagnosis of cancerous tissues and organs. In this paper, we intend to systematically review the state-of-the-art research on AI-based ML and DL techniques applied to the modeling of colorectal cancer. All research papers in the field of colorectal cancer are collected based on ML and DL techniques, and they are then classified into three categories: the aim of the prediction, the method of the prediction, and data samples. Following that, a thorough summary and a list of the studies gathered under each topic are provided. We conclude our study with a critical discussion of the challenges and opportunities in colorectal cancer prediction using ML and DL techniques by concentrating on the technical and medical points of view. Finally, we believe that our study will be helpful to scientists who are considering employing ML and DL methods to diagnose colorectal cancer.
... Feature selection methods are an essential part in the classification models. The reduced number of features is used in classification tasks to enhance the performance of the model (Rahman and Muniyandi 2018;Sahran et al. 2018). A work implementing genetic algorithm-based feature selection technique has been developed by Alickovic and Subasi (2017) and Ahmad et al. (2015). ...
Breast cancer is a life-threatening and consequential disease due to its invasive and proliferative trait, predominantly found in women. Early detection of the cancer is a significant contributor to improved mortality and hence is an area of keen focus for ongoing researches. However, developing a technique to diagnose the severity of the patients at an early stage is a challenging task. Manual diagnostic techniques are time-consuming and result in inaccurate diagnosis of breast cancer. Prompted by these facts, a quantum optimized rule-base generated automated framework is developed to cluster the data based on degree of criticality of the cancer patients and further classify it as benign or malignant utilizing probability of malignancy of the clusters along with assignment of grades of cancer. Firstly, after implementing data pre-processing step, significant features are selected using an integrated feature selection approach. An efficient weightage algorithm is proposed incorporating the knowledge of physicians and the benefits of regression analysis which thereby provides a novel approach for detection of breast cancer. A novel ensemble clustering and classification algorithm employing voting-based Weighted Interval Type-II Fuzzy Inference System and Staged Pegasos Quantum Support Vector Classifier is then developed basis the prioritization of clusters depicting the critical state of breast cancer. A grading approach is also proposed based on fuzzy linguistic multi-criteria decision making system. Finally, the research is validated on Wisconsin Breast Cancer dataset. The detailed implementation of the proposed integrated model is accomplished to establish its superiority over other existing models in the literature.
... GA is wrapped with Naïve Bayes classifier for gene selection of diabetes data [10]. Breast First Search (BFS) and very well-known learning algorithm Artificial Neural Network (ANN) are used to select features and classify colon data [11]. The metaheuristic Bat Algorithm (BA) is embedded with Optimum Path Forest (OPF) algorithm to extract key features of medical data [12]. ...
Classification of different deadly diseases using machine learning algorithms helps in making the health care system more robust. This, not only reduces the human errors during diagnosis of disease due to inexperience but also helps the physician for taking an emergency action earlier to biopsy. As the genomic data undergoes through the malediction of excessive dimension problem, both the selection of the remarkable genes and classification of these data efficiently still remain as a demanding research problem. To obtain the notable features from high dimensional data, by using a nature-inspired algorithm, is a Non-deterministic Polynomial-time (NP)-Hard problem. Therefore, a researcher can apply new algorithm to solve this issue. In this suggested approach, an integrated natured-inspired algorithm i.e., Sine-Cosine (SC) based Monarch Butterfly Optimization (SC-MBO) is merged with an Adaptive Broad Learning System (ABLS) called SC-MBO-ABLS, to find out the most notable genetic features and classify the genomic data at the same time. In the preliminary stage, a feature extraction method i.e. Kernel based Fisher Score (K-FS) is applied to extract a notable key gene subset. Then, this extracted key gene subset goes through further execution with the SC-MBO-ABLS method. To examine the effectiveness of the presented method, ten genomic datasets are considered. Here, several performance evaluators (i.e., Precision, Matthews Correlation Coefficient, Sensitivity, Kappa, Specificity, and F-score) are used for a neutral estimation of the presented approach. This presented model is compared with SC-MBO wrapped Multilayer Perceptron (SC-MBO-MLP), SC-MBO wrapped Extreme Learning Machine (SC-MBO-ELM), and SC-MBO wrapped Kernel Extreme Learning Machine (SC-MBO-KELM). Further, sixteen existing standard models are compared to prove the supremacy of the presented method. Moreover, an analysis of Variance i.e., ANOVA test is carried out for statistical evaluation of the proposed work. Eventually, according to the above quantitative and qualitative measure, it is summarized that the suggested SC-MBO-ABLS method surpasses other considering standard models.
... F1-Score is the number of categories that can be predicted by the system correctly, with mathematical formulas that can be calculated using equation (7). The mathematical formulation of the confusion matrix is [22]. ...
This study proposes the MLP type ANN method with the influence of Cross-Validation evaluation using three K-Fold tests, namely 3, 5, and 8. The data used are the data that relate to COVID-19 cases issued from the Bandung Public Health Office, climate report from Bandung Meteorological, Climatological, and Geophysical Agency (BMKG), population data from Bandung Population Office, citizen’s educational history from the Bandung Education Office and the West Java Open Data website. The data also was gathered from 151 sub-districts in Bandung City, with a total of 22 attributes collected from November 2020 to December 2021. The ANN method is included in the deep learning process. Therefore, the number of hidden layers utilized has a significant impact on the performance of the model being constructed. The implementation of Cross-Validation evaluation with K=8 results in an accuracy value of up to 98% and an error metric measurement of 0.3404 for MAE and 0.5994 for RMSE. This study's objective is to provide information on the optimal K-Fold Cross Validation parameters used in this ANN method to provide better performance during building a classification model for confirmed Covid-19 patients.
... Feature selection methods are an essential part in the classification models. The reduced number of features are used in classification tasks to enhance the performance of the model [18,19]. A work implementing genetic algorithm based feature selection technique, has been developed by Aličković and Subasi [20] and Ahmad et al. [21]. ...
Breast cancer is a major threat, predominantly affecting the female population. Staging of cancer enables early detection and prognosis of patients, leading to determination of efficient and accurate treatment. Consequently, simplified models are required to integrate heterogeneous data for deriving knowledge about patients for further treatment. To achieve this goal, developing machine learning based diagnostic techniques is the predominant need. Prompted by these facts, a novel diagnostic model for staging of breast cancer infusing ensemble clustering, feature weighting based ranking of clusters and ensemble classification into benign or malignant class is developed. The proposed work constitutes of five different phases: data pre-processing, feature selection, ensemble clustering, ensemble classification, and staging of cancer. This work first employs Multiple Imputation Chained Equation for imputing missing values, followed by proposed feature selection technique employing Association Rules, Classification and Regression Tree, and Fuzzy Logic. Subsequently, a coupled clustering and classification algorithm based on consensus is developed to cluster features from different datasets using Self-Organizing Map and Decision Tree. A hierarchical clustering based ranking of these clusters using Multilinear Regression and Modified Fuzzy Analytical Hierarchical Process is proposed to prioritize features. Next, a staged classifier is developed integrating Probabilistic Fuzzy Logic and Multilayer Perceptron followed by feature extraction based staging of cancer. Finally, proposed work is validated on four datasets with various performance metrics using different combinations of train-test dataset. Moreover, k-fold cross-validation is implemented to eliminate biasedness. The detailed analysis of results of this work showcases superiority over other state-of-art methods in literature.
... The accuracy results obtained are 89.52%. Research [11] apply the ANN algorithm in the medical field, especially in cancer research. While research [10] on the prediction of monthly weather classification. ...
... Parameters evaluated were accuracy, precision, recall, and f1-score. Accuracy is the ability of the model to predict the correct class, the accuracy formula is computable using an equation (11). Precision is the ratio of the class of rainfall that is predicted correctly, and the precision formulation can be calculated using equation (12). ...
This study proposes several methods to analyze the performance of the hybrid machine learning method using Voting and Stacking on rainfall classification. The two hybrid methods will combine five classification methods, namely Logistic Regression, Support Vector Machine, Random Forest, Artificial Neural Network, and eXtreme Gradient Boosting. The data used is Bandung City rainfall data for the years 2005 until 2021. The hybrid method is classified as an ensemble, which means combining several individual classification models to improve the performance of the built model. Voting algorithm has weaknesses in imbalanced data, while stacking does not. The results show that by combining five machine learning methods on an imbalanced dataset, the Stacking algorithm obtains an accuracy value of 99.60%. Meanwhile, with the addition of the SMOTE technique, the accuracy increases to 99.71%. This is supported by the performance of the Stacking method which is superior because it takes the best classification value for each individual model and can overcome the imbalance. Model evaluation does not only focus on accuracy, but also precision, recall, and f1-score. The contribution of this research is to provide information about the best Hybrid method between Voting and Stacking in obtaining model performance results on rainfall classification.
... GA is applied for gene selection and Naïve Bayes is considered as a classifier for Indian diabetes data [10]. Best First Search technique is wrapped with an Artificial Neural Network learning algorithm for feature selection and classification of colon cancer data [11]. Bat algorithm (BA) is wrapped with optimum path forest (OPF) learning algorithm for feature selection of biomedical data [12]. ...
... Here, Z max = maximum walk step of individual butterfly in each step and t is the current generation and dY n represents the walk step of the individual butterfly i which can be formulated by taking the concept of Levy flight method as (11). ...
The classification of high dimensional gene expression/ microarray data always plays an important role in various disease diagnoses and drug discovery. To avoid the curse of high dimensionality, the selection of the most influential genes remains a challenging task for the researchers in the machine learning field. As the extraction of influential features by a bio-inspired algorithm is a non-deterministic polynomial-time (NP-Hard) task, the possibility of applying new algorithm is always there. In this suggested work, a recently developed bio-inspired algorithm, Monarch Butterfly Optimization (MBO), is wrapped with the Broad Learning System (BLS), called MBO-BLS, to choose the most influential features and classify the microarray data simultaneously. In the first stage, a pre-selection method (Relief) is used to select a feature subset. Then, this selected feature subset undergoes further execution with the MBO-BLS model. To estimate the efficacy of the presented model, six cancerous microarray datasets are taken. Here, sensitivity, specificity, precision, F-score, Kappa, and MCC measures are used for an impartial comparison. Further, to prove the supremacy of the presented method, the basic BLS, Genetic Algorithm wrapped BLS (GA-BLS), Particle Swarm Optimization wrapped BLS (PSO-BLS), and the existing ten models are taken for comparison. Moreover, to examine the designed model statistically, Analysis of variance (ANOVA) test is also performed here. From the above qualitative and quantitative analysis, it is concluded that the proposed MBO-BLS model outclasses other considering models.
... In this paper, HOFS provides the most valuable 38 feature vectors which are chosen for classification. This part makes centre around acquiring a subset of F that can accomplish better execution of classification [31], [32]. In this paper, mammogram images are labeled as normal or abnormal lesion. ...
p> Breast cancer is the most driving reason for death in women in both developed and developing nations. For the plan of effective classification of a system, the selection of features method must be used to decrease irregularity part in mammogram images. The proposed approach is used to crop the region of interests (ROIs) manually. Based on that number of features are extracted. In this proposed method a novel hybrid optimum feature selection (HOFS) method is used to find out the significant features to reach maximum accuracy for this classification. A number of selected features is applied to train the neural network. In this proposed method accessible informational index from the mini–mammographic image analysis society (MIAS) database was used. The classification of this mammogram database involved a neural networks classifier which attained an accuracy of 99.7% with a sensitivity of 99.5%, and specificity of 100% as the area under the curve (AUC) is 0.9975 and matthew’s correlation coefficient (MCC) represents a binary class value which reached the value of 0.9931. It can be useful in a computer-aided diagnosis system (CAD) framework to help the radiologist in analyzing breast cancer. Results achieved with the proposed method are better compared to recent work. </p
... In [13], the authors proposed an ANN for the classification of cancer and feature selection of the colon cancer dataset. For feature selection, the best first search technique was used in the Weka tools and the accuracy achieved was 98.4% and without feature selection, the accuracy obtained was 95.2%. ...
Colorectal cancer is one major cause of cancer-related death around the globe. Recent breakthroughs in deep learning have paved the way to apply it for the automation of histopathology images as a tool for computer-aided diagnosis of medical imaging. Here we have presented a novel state of the art classification model for classifying the colorectal histopathology images into 9 classes. All the traditional approaches like texture-based classification, transfer learning etc. already has been used to achieve a state-of-the-art result, but these have some limitations. Rather than using conventional mechanisms, we have proposed a methodology that can interpret the histopathology images in a more generalized way without image preprocessing and augmentation mechanisms. A combination of two deep learning architectures i.e., an encoder unit of autoencoder module and a modified DenseNet121 architecture are used for this purpose. An accuracy of 97.2% on Zenodo 100k colorectal histopathology dataset has been reported. The presented result is better than most of the contemporary works in this domain. We have also evaluated the effectiveness of the current approach for the low-resolution histopathological images and achieved good recognition accuracy.
... The purpose of dimensionality reduction is to obtain optimal features rapidly, since the more the data dimensionality is, the more the optimization of features will get slower. So, feature selection's primary goal is to ascertain only minor features from a specific problem domain that signifies high-ranking classification performance [8,9]. ...
Autism Spectrum Disorder (ASD), according to DSM-5 in the American Psychiatric Association, is a neurodevelopmental disorder that includes deficits of social communication and social interaction with the presence of restricted and repetitive behaviors. Children with ASD have difficulties in joint attention and social reciprocity, using non-verbal and verbal behavior for communication. Due to these deficits, children with autism are often socially isolated. Researchers have emphasized the importance of early identification and early intervention to improve the level of functioning in language, communication, and well-being of children with autism. However, due to limited local assessment tools to diagnose these children, limited speech-language therapy services in rural areas, etc., these children do not get the rehabilitation they need until they get into compulsory schooling at the age of seven years old. Hence, efficient approaches towards early identification and intervention through speedy diagnostic procedures for ASD are required. In recent years, advanced technologies like machine learning have been used to analyze and investigate ASD to improve diagnostic accuracy, time, and quality without complexity. These machine learning methods include artificial neural networks, support vector machines, a priori algorithms, and decision trees, most of which have been applied to datasets connected with autism to construct predictive models. Meanwhile, the selection of features remains an essential task before developing a predictive model for ASD classification. This review mainly investigates and analyzes up-to-date studies on machine learning methods for feature selection and classification of ASD. We recommend methods to enhance machine learning's speedy execution for processing complex data for conceptualization and implementation in ASD diagnostic research. This study can significantly benefit future research in autism using a machine learning approach for feature selection, classification, and processing imbalanced data.
