Gene selection stability's dependence on dataset difficulty

Robustness or stability of feature selection techniques is a topic of recent interest, and is an important issue when selected feature subsets are subsequently analysed by domain experts to gain more insight into the problem modelled. In this work, we investigate the use of ensemble feature selection techniques, where multiple feature selection methods are combined to yield more robust results. We show that these techniques show great promise for high-dimensional domains with small sample sizes, and provide more robust feature subsets than a single feature selection technique. In addition, we also investigate the effect of ensemble feature selection techniques on classification performance, giving rise to a new model selection strategy.

Ensemble gene (feature) selection is a promising new strategy with many benefits including more stable gene lists and improved classification results. The ensemble portion is achieved through multiple runs of feature selection which are then aggregated into a single result. The critical question is how many iterations of feature selection are appropriate. Too few iterations can make classification performance suffer. However, too many iterations will cause issues regarding computational costs. The goal is to choose the correct number of iterations to maximize classification performance without expending too much computational power. Our paper is an in-depth study on the effect of the number of iterations of feature selection on classification performance. Our work employs eleven DNA microarray datasets, on which we apply various ensemble methods, feature selection techniques, classifiers, and feature subset sizes. The results show that using 10 iterations of feature ranking during ensemble feature selection is not sufficient to optimize classification results and that a larger number of iterations is required (20 or 50). However, there is very little distinction between 20 iterations and 50 iterations, as both produce very similar classification results. Our recommendation is to use 20 iterations because while 20 iterations and 50 iterations perform similarly, 20 iterations has a much smaller computation time. To our knowledge there has not been a previous study as expansive as this one on the effects of the number of iterations of feature selection on ensemble feature selection.

Text mining is a special case of data mining which explore unstructured or semi-structured text documents, to establish valuable patterns and rules that indicate trends and significant features about specific topics. Text mining has been in pattern recognition, predictive studies, sentiment analysis and statistical theories in many areas of research, medicine, financial analysis, social life analysis, and business intelligence. Text mining uses concept of natural language processing and machine learning. Machine learning algorithms have been used and reported to give great results, but their performance of machine learning algorithms is affected by factors such as dataset domain, number of classes, length of the corpus, and feature selection techniques used. Redundant attribute affects the performance of the classification algorithm, but this can be reduced by using different feature selection techniques and dimensionality reduction techniques. Feature selection is a data preprocessing step that chooses a subset of input variable while eliminating features with little or no predictive information. Feature selection techniques are Information gain, Term Frequency, Term Frequency-Inverse document frequency, Mutual Information, and Chi-Square, which can use a filters, wrappers, or embedded approaches. To get the most value from machine learning, pairing the best algorithms with the right tools and processes is necessary. Little research has been done on the effect of feature selection techniques on classification accuracy for pairing of these algorithms with the best feature selection techniques for optimal results. In this research, a text classification experiment was conducted using incident management dataset, where incidents were classified into their resolver groups. Support vector machine (SVM), K-Nearest Neighbors (KNN), Naïve Bayes (NB) and Decision tree (DT) machine learning algorithms were examined. Filtering approach was used on the feature selection techniques, with different ranking indices applied for optimal feature set and classification accuracy results analyzed. The classification accuracy results obtained using TF were, 88% for SVM, 70% for NB, 79% for Decision tree, and KNN had 55%, while Boolean registered 90%, 83%, 82% and 75%, for SVM, NB, DT, and KNN respectively. TF-IDF, had 91%, 83%, 76%, and 56% for SVM, NB, DT, and KNN respectively. The results showed that algorithm performance is affected by feature selection technique applied. SVM performed best, followed by DT, KNN and finally NB. In conclusion, presence of noisy data leads to poor learning performance and increases the computational time. The classifiers performed differently depending on the feature selection technique applied. For optimal results, the classifier that performed best together with the feature selection technique with the best feature subset should be applied for all types of data for accurate classification performance. Keywords: Text Classification, Supervised Machine Learning, Feature Selection DOI : 10.7176/JIEA/9-3-06 Publication date :May 31 st 2019

Due to the prevalence of high dimensionality(having a large number of independent attributes), feature selection techniques (which reduce the feature subset to amore manageable size) have become quite popular. These reduced feature subsets can help improve the performance of classification models and can also inform researchers about which features are most relevant for the problem at hand. For this latter problem, it is often most important that the features chosen are consistent even in the face of changes(perturbations) to the dataset. While previous studies have considered the problem of finding so-called "stable" feature selection techniques, none has examined stability across all three major categories of feature selection technique: filter-based feature rankers (which use statistical measures to assign scores to each feature), filter-based subset evaluators (which also employ statistical approaches, but consider whole feature subsets at a time), and wrapper-based subset evaluation (which also considers whole subsets, but which builds classification models to evaluate these subsets). In the present study, we use two datasets from the domain of Twitter profile mining to compare the stability of five filter-based rankers, two filter-based subset evaluators, and five wrapper-based subset evaluators. We find that the rankers are most stable, followed by the filter-based subset evaluators, with the wrappers being the least stable. We also show that the relative performance among the techniques within each group is consistent across dataset and perturbation level. However, the relative stability of the two datasets does vary between the groups, showing that the effects are more complex than simply "one group is always more stable than another group".

<p>Classification aims to identify a class label of an instance according to the information from its characteristics or features. Unfortunately, many classification problems have a large feature set containing irrelevant and redundant features, which reduce the classification performance. In order to address the above problem, feature selection is proposed to select a small subset of relevant features. There are three main types of feature selection methods, i.e. wrapper, embedded and filter approaches. Wrappers use a classification algorithm to evaluate candidate feature subsets. In embedded approaches, the selection process is embedded in the training process of a classification algorithm. Different from the other two approaches, filters do not involve any classification algorithm during the selection process. Feature selection is an important process but it is not an easy task due to its large search space and complex feature interactions. Because of the potential global search ability, Evolutionary Computation (EC), especially Particle Swarm Optimization (PSO), has been widely and successfully applied to feature selection. However, there is potential to improve the effectiveness and efficiency of EC-based feature selection. The overall goal of this thesis is to investigate and improve the capability of EC for feature selection to select small feature subsets while maintaining or even improving the classification performance compared to using all features. Different aspects of feature selection are considered in this thesis such as the number of objectives (single-objective/multi-objective), the fitness function (filter/wrapper), and the searching mechanism. This thesis introduces a new fitness function based on mutual information which is calculated by an estimation approach instead of the traditional counting approach. Results show that the estimation approach works well on both continuous and discrete data. More importantly, mutual information calculated by the estimation approach can capture feature interactions better than the traditional counting approach. This thesis develops a novel binary PSO algorithm, which is the first work to redefine some core concepts of PSO such as velocity and momentum to suit the characteristics of binary search spaces. Experimental results show that the proposed binary PSO algorithm evolve better solutions than other binary EC algorithms when the search spaces are large and complex. Specifically, on feature selection, the proposed binary PSO algorithm can select smaller feature subsets with similar or better classification accuracies, especially when there are a large number of features. This thesis proposes surrogate models for wrapper-based feature selection. The surrogate models use surrogate training sets which are subsets of informative instances selected from the training set. Experimental results show that the proposed surrogate models assist PSO to reduce the computational cost while maintaining or even improving the classification performance compared to using only the original training set. The thesis develops the first wrapper-based multi-objective feature selection algorithm using MOEA/D. A new decomposition strategy using multiple reference points for MOEA/D is designed, which can deal with different characteristics of multi-objective feature selection such as highly discontinuous Pareto fronts and complex relationships between objectives. The experimental results show that the proposed algorithm can evolve more diverse non-dominated sets than other multi-objective algorithms. This thesis introduces the first PSO-based feature selection algorithm for transfer learning. In the proposed algorithm, the fitness function uses classification performance to reduce the differences between domains while maintaining the discriminative ability on the target domain. The experimental results show that the proposed algorithm can select feature subsets which achieve better classification performance than four state-of-the-art feature-based transfer learning algorithms.</p>

Text clustering involves data that are of very high dimension. Feature selection techniques find subsets of relevant features from the original feature space that help in efficient and effective clustering. Selection of relevant features merely on ranking scores without considering correlation interferes with the clustering performance. An efficient feature selection technique should be capable of preserving the multi-cluster structure of the data. The purpose of the present work is to demonstrate that feature selection techniques which take into consideration the correlation among features in multi-cluster scenario show better clustering results than those techniques that simply rank features independent of each other. This paper compares two feature selection techniques in this regard viz. the traditional Tf-Idf and the Multi-Cluster Feature Selection (MCFS) technique. The experimental results over the TDT2 and Reuters-21,578 datasets show the superior clustering results of MCFS over traditional Tf-Idf.

Computer-aided diagnosis has the potential of increasing diagnostic accuracy by providing a second reading to radiologists. In many computerized schemes, numerous features can be extracted to describe suspect image regions. A subset of these features is then employed in a data classifier to determine whether the suspect region is abnormal or normal. Different subsets of features will, in general, result in different classification performances. A feature selection method is often used to determine an "optimal" subset of features to use with a particular classifier. A classifier performance measure (such as the area under the receiver operating characteristic curve) must be incorporated into this feature selection process. With limited datasets, however, there is a distribution in the classifier performance measure for a given classifier and subset of features. In this paper, we investigate the variation in the selected subset of "optimal" features as compared with the true optimal subset of features caused by this distribution of classifier performance. We consider examples in which the probability that the optimal subset of features is selected can be analytically computed. We show the dependence of this probability on the dataset sample size, the total number of features from which to select, the number of features selected, and the performance of the true optimal subset. Once a subset of features has been selected, the parameters of the data classifier must be determined. We show that, with limited datasets and/or a large number of features from which to choose, bias is introduced if the classifier parameters are determined using the same data that were employed to select the "optimal" subset of features.

Introduction: The early and accurate detection of COVID-19 remains a life-threatening challenge in medical analysis. Machine learning is used for predicting disease outcomes based on clinical parameters. This analysis proposes a comparative analysis of feature selection method and classification techniques to enhance COVID-19 detection accuracy using blood biomarkers. We used a pensourse dataset of 1,724 cases, including 35 features. To improve the model performance data preprocessing process included outlier handling, normalization, and transformation techniques to improve model performance. To identify the relevant features, we employed the three-feature selection methods Chi-Square test, Pearson correlation coefficient, and Random Forest. The model prediction accuracy was enhanced using a stacking ensemble classication techniques. The machine learning based classification models effectively predicted COVID-19 infectious disease using blood biomarkers with optimized feature selection techniques. Objectives: To enhance the accuracy of COVID-19 prediction using machine learning techniques by applying feature selection and classification techniques on blood biomarkers. Methods: The comparative analysis utilized a publicly available dataset containing 1,724 cases with 35 attributes. Data preprocessing involved outlier handling, normalization, and transformation techniques. Employed Chi-Square test, Pearson correlation coefficient, and Random Forest feature selection techniques. Stacking ensemble classification algorithm was utilized for the better performance of a model. Results: The classification models demonstrated efficiency in predicting COVID-19 using blood biomarkers. Optimized feature selection significantly improved predictive accuracy, highlighting the importance of selecting relevant features for model performance enhancement. Conclusions: This study showcases the potential of ML-driven approaches for COVID-19 detection, emphasizing the role of feature selection in improving classification accuracy. The findings contribute to the advancement of diagnostic tools, offering a data-driven solution for rapid and reliable COVID-19 screening.

Credit risk assessment is an essential task for financial institutions, as it evaluates the likelihood of loan defaults. Misclassifying loan applications can result in substantial economic losses. However, accurately predicting customers' creditworthiness presents challenges for machine learning models due to high dimensionality and imbalanced class distributions, which can severely impact overall model performance. This implements three feature selection (FS) techniques, such as Random Forest with Highly Correlated features (RFHC), Information Gain, and Chi-Square to address the high dimensional problem. Additionally, SMOTE is implemented to address the imbalanced dataset. The model performance is analyzed using three different approaches. First, employing only feature selection techniques, in the second approach, we employed feature selection followed by oversampling, and in the final approach, oversampling prior to feature selection is employed. The experimental results indicate that implementing feature selection followed by oversampling gives better performance compared to the other two approaches. Among the feature selection techniques, RFHC outperforms the others, achieving a highest F1-score of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{95.12 \%}$</tex> and an AUC value of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{85.76 \%}$</tex>.

Last century, the challenge was to develop new technologies that store large amount of data. Recently, the challenges are to effectively utilize the incredible amount of data and to obtain knowledge that benefits business, scientific, and government transactions by using subset of features rather than the whole features in the dataset. In the present paper, we have focused on feature selection techniques as a method to gain high quality attributes to enhance the mining process. Feature selection techniques touch all disciplines that require knowledge discovery from large data. In our study, we made a comparison between benchmark feature selection methods based on three well-known datasets and three well-recognized machine learning algorithms. The study found that feature selection methods are capable to improve the performance of learning algorithms. However, no single feature selection methods that best satisfy all datasets and learning algorithms. Therefore, machine learning researchers should understand the nature of datasets and learning algorithms characteristics in order to obtain better outcome. Overall, correlation based feature selection (CFS) and consistency based subset evaluation (CB) performed better than information gain, symmetrical uncertainty, Relief (RF), and principle components analysis (PC).

&lt;p&gt;Classification problems often have a large number of features, but not all of them are useful for classification. Irrelevant and redundant features may even reduce the classification accuracy. Feature selection is a process of selecting a subset of relevant features, which can decrease the dimensionality, shorten the running time, and/or improve the classification accuracy. There are two types of feature selection approaches, i.e. wrapper and filter approaches. Their main difference is that wrappers use a classification algorithm to evaluate the goodness of the features during the feature selection process while filters are independent of any classification algorithm. Feature selection is a difficult task because of feature interactions and the large search space. Existing feature selection methods suffer from different problems, such as stagnation in local optima and high computational cost. Evolutionary computation (EC) techniques are well-known global search algorithms. Particle swarm optimisation (PSO) is an EC technique that is computationally less expensive and can converge faster than other methods. PSO has been successfully applied to many areas, but its potential for feature selection has not been fully investigated. The overall goal of this thesis is to investigate and improve the capability of PSO for feature selection to select a smaller number of features and achieve similar or better classification performance than using all features. This thesis investigates the use of PSO for both wrapper and filter, and for both single objective and multi-objective feature selection, and also investigates the differences between wrappers and filters. This thesis proposes a new PSO based wrapper, single objective feature selection approach by developing new initialisation and updating mechanisms. The results show that by considering the number of features in the initialisation and updating procedures, the new algorithm can improve the classification performance, reduce the number of features and decrease computational time. This thesis develops the first PSO based wrapper multi-objective feature selection approach, which aims to maximise the classification accuracy and simultaneously minimise the number of features. The results show that the proposed multi-objective algorithm can obtain more and better feature subsets than single objective algorithms, and outperform other well-known EC based multi-objective feature selection algorithms. This thesis develops a filter, single objective feature selection approach based on PSO and information theory. Two measures are proposed to evaluate the relevance of the selected features based on each pair of features and a group of features, respectively. The results show that PSO and information based algorithms can successfully address feature selection tasks. The group based method achieves higher classification accuracies, but the pair based method is faster and selects smaller feature subsets. This thesis proposes the first PSO based multi-objective filter feature selection approach using information based measures. This work is also the first work using other two well-known multi-objective EC algorithms in filter feature selection, which are also used to compare the performance of the PSO based approach. The results show that the PSO based multiobjective filter approach can successfully address feature selection problems, outperform single objective filter algorithms and achieve better classification performance than other multi-objective algorithms. This thesis investigates the difference between wrapper and filter approaches in terms of the classification performance and computational time, and also examines the generality of wrappers. The results show that wrappers generally achieve better or similar classification performance than filters, but do not always need longer computational time than filters. The results also show that wrappers built with simple classification algorithms can be general to other classification algorithms.&lt;/p&gt;

In this work, we propose a novel Feature Selection framework, called Sparse-Modeling Based Approach for Class Specific Feature Selection (SMBA-CSFS), that simultaneously exploits the idea of Sparse Modeling and Class-Specific Feature Selection. Feature selection plays a key role in several fields (e.g., computational biology), making it possible to treat models with fewer variables which, in turn, are easier to explain, by providing valuable insights on the importance of their role, and might speed the experimental validation up. Unfortunately, also corroborated by the no free lunch theorems, none of the approaches in literature is the most apt to detect the optimal feature subset for building a final model, thus it still represents a challenge. The proposed feature selection procedure conceives a two steps approach: (a) a sparse modeling-based learning technique is first used to find the best subset of features, for each class of a training set; (b) the discovered feature subsets are then fed to a class-specific feature selection scheme, in order to assess the effectiveness of the selected features in classification tasks. To this end, an ensemble of classifiers is built, where each classifier is trained on its own feature subset discovered in the previous phase, and a proper decision rule is adopted to compute the ensemble responses. In order to evaluate the performance of the proposed method, extensive experiments have been performed on publicly available datasets, in particular belonging to the computational biology field where feature selection is indispensable: the acute lymphoblastic leukemia and acute myeloid leukemia, the human carcinomas, the human lung carcinomas, the diffuse large B-cell lymphoma, and the malignant glioma. SMBA-CSFS is able to identify/retrieve the most representative features that maximize the classification accuracy. With top 20 and 80 features, SMBA-CSFS exhibits a promising performance when compared to its competitors from literature, on all considered datasets, especially those with a higher number of features. Experiments show that the proposed approach might outperform the state-of-the-art methods when the number of features is high. For this reason, the introduced approach proposes itself for selection and classification of data with a large number of features and classes.

Today, video data, as a powerful multimedia component, is accompanied by some problems with increasing usage in communication, health, education, and social media in particular. Classification and detection of concepts in video data by automatic methods are some of these challenging problems. In this study, we propose a video classification system, which incorporates deep convolutional neural networks (CNNs) by leveraging feature selection and data fusion techniques to improve the accuracy of the classification. Principal Component Analysis (PCA) as a feature selection method and Discriminant Correlation Analysis (DCA) technique, which incorporates class associations into the correlation analysis of feature sets for data fusion, are applied to the problem at the feature level. Support Vector Machines (SVMs) have been trained with new feature vectors obtained from different deep convolutional neural networks by feature selection and data fusion methods. The proposed method is tested for 38 concepts on TRECVID 2013 SIN video task dataset and the results are evaluated. Our results show that the classification accuracy is improved by 4% with an accuracy of 50.27% when the proposed data fusion and feature selection techniques are used.

In the domain of bioinformatics, two common problems encountered when analyzing real-world datasets are class imbalance and high dimensionality. Boosting is a technique that can be used to improve classification performance, even in the presence of class imbalance. In addition, data sampling and feature selection are two important preprocessing techniques used to counter the adverse effects of both challenges collectively. In this study, we examine whether the inclusion of boosting along with joint deployment of feature selection and data sampling techniques affect the classification performance of inductive models. To this end, we used two approaches: filter-based feature selection followed by either data sampling (denoted as FS-DS) or a hybrid data sampling and boosting technique entitled RUSBoost (denoted as FRB) which integrates random under sampling within the boosting process. We conducted an extensive experimental study using six high dimensional and imbalanced bioinformatics datasets along with three learners and four feature subset sizes. Our results show that the improvement of classification performance due to boosting depends on the choice of learner used to build the model. We recommend FRB because it outperforms FS-DS for nearly all scenarios. Additionally, our ANOVA analysis shows that the FRB is statistically distinguishable from the FS-DS when using the LR learner. To our knowledge, this is the first study to investigate the effects of boosting along with combined feature selection and data sampling on classification performance of inductive models in the domain of bioinformatics.

Innovations in machine learning algorithms have enhanced the effectiveness of malware detection systems during the previous decades. However, the advancement of high throughput technologies results in high dimensional malware data, making feature selection useful and mandatory in such datasets. The feature selection technique is an information retrieval tool that aims to improve classifiers by listing important features, which also aids in reducing computational overload. However, different feature selection algorithms select representative features using various criteria, making it difficult to determine the optimal technique for distinct domain datasets. Ensemble feature selection approaches, which integrate the results of several feature selections, can be used to overcome the inadequacies of single-feature selection methods. Therefore, this paper attempts to determine whether the heterogeneous ensemble of filter and embedded feature selection approaches, namely the heterogenous ensemble of ANOVA F-test, ReliefF, L1-penalized logistic regression, LASSO regression, Extra-Tree Classifier and XGBoost feature selection techniques, namely HEFS-ARLLEX, can provide a better classification performance than the single feature selection technique and other ensemble feature selection approaches for malware classification data. The experimental results show that HEFS-ARLLEX, which combines both filters and embedded, is a better choice, providing consistently high classification accuracy, recall, precision, specificity and F-measure and a reasonable feature reduction rate for malware classification dataset.