Greedy Stepwise Search Research Articles

Multi-view feature fusion and density-based minority over-sampling technique for amyloid protein prediction under imbalanced data

Amyloid fibers formed by the aggregation of amyloid proteins can lead to various neurological diseases. Accurate prediction of amyloid proteins can provide scientific basis for exploring the pathogenesis of related neurological diseases, and developing targeted treatment plans. In view of the limitations of the synthetic minority over-sampling technique (SMOTE), an improved SMOTE method based on density clustering is proposed in this paper, and combined with multi-view features to construct an amyloid protein prediction model. The benefits of this enhanced SMOTE method include eliminating problematic samples, accomplishing an evenly distributed minority sample pool, and utilizing all available boundary sample information. Compared with other over-sampling methods, this method can effectively improve the authenticity and representativeness of the synthesized samples, and has a strong ability to process imbalanced data. After the dataset is over-sampling, different prediction algorithms are trained respectively to construct the baseline models. To fully characterize proteins, the probability and category features are extracted from the baseline models and combined with the sequence features selected by CFS-GSSS (Correlation-based Feature Selection Combined with Greedy Stepwise Search Strategy) to generate the multi-view features. Experimental results demonstrate the complementarity and feasibility of the multi-view features. On the independent test set, compared with the existing best model (ECAmyloid), the proposed amyloid protein prediction model improved the sensitivity and geometric mean by 0.0706 and 0.0129, respectively, far superior to other existing methods.

Modelling Occurrence of Invasive Water Hyacinth (Eichhornia crassipes) in Wetlands

Eichhornia crassipes is one of the most ubiquitous invasive aquatic species in the world that negatively impact on native fauna and flora. The information on the occurrence of such invasive plant is crucial for the river/wetland management. The aim of the present study is to apply classification tree model (testing with the highest level of pruning confidence factor) integrated with an optimizer technique (greedy stepwise search algorithm) to predict the occurrence of this exotic species based on water quality and physical-habitat parameters. The ten sites (in the Anzali wetland, located in northern Iran) were monthly measured where the exotic species was present in 50 % of sampling sites and it was absent in the remaining of the sites. In total, 120 samples of E. crassipes (60 presence and 60 absence instances) were monthly measured together with 12 environmental variables during 1-year study period (2017–2018). For the model, two-third of datasets (80 instances) was employed for the training and the remainder for the validation set (40 instances). Before model optimizing (with the pruning confidence factor, PCF = 0.01), six variables were predicted by the model in three folds confirming that the non-occurrence of the exotic species might be associated with increasing flow velocity, depth of ecosystem, water turbidity, bicarbonate and dissolved oxygen concentration, while the occurrence of the exotic species (in terms of the abundance) might show an increase with increasing the concentration of nutrients such as phosphate. After model optimizing (with PCF = 0.01), the model selected five variables in three folds (flow velocity, depth, phosphate, bicarbonate, and nitrate) so that except nitrate, other selected variables were in common before and after variable selection. CT integrated with GS model thus proved to have a high potential when applied for decision-making in the context of wetland management.

Performance Evaluation of a Proposed Machine Learning Model for Chronic Disease Datasets Using an Integrated Attribute Evaluator and an Improved Decision Tree Classifier

There is a consistent rise in chronic diseases worldwide. These diseases decrease immunity and the quality of daily life. The treatment of these disorders is a challenging task for medical professionals. Dimensionality reduction techniques make it possible to handle big data samples, providing decision support in relation to chronic diseases. These datasets contain a series of symptoms that are used in disease prediction. The presence of redundant and irrelevant symptoms in the datasets should be identified and removed using feature selection techniques to improve classification accuracy. Therefore, the main contribution of this paper is a comparative analysis of the impact of wrapper and filter selection methods on classification performance. The filter methods that have been considered include the Correlation Feature Selection (CFS) method, the Information Gain (IG) method and the Chi-Square (CS) method. The wrapper methods that have been considered include the Best First Search (BFS) method, the Linear Forward Selection (LFS) method and the Greedy Step Wise Search (GSS) method. A Decision Tree algorithm has been used as a classifier for this analysis and is implemented through the WEKA tool. An attribute significance analysis has been performed on the diabetes, breast cancer and heart disease datasets used in the study. It was observed that the CFS method outperformed other filter methods concerning the accuracy rate and execution time. The accuracy rate using the CFS method on the datasets for heart disease, diabetes, breast cancer was 93.8%, 89.5% and 96.8% respectively. Moreover, latency delays of 1.08 s, 1.02 s and 1.01 s were noted using the same method for the respective datasets. Among wrapper methods, BFS’ performance was impressive in comparison to other methods. Maximum accuracy of 94.7%, 95.8% and 96.8% were achieved on the datasets for heart disease, diabetes and breast cancer respectively. Latency delays of 1.42 s, 1.44 s and 132 s were recorded using the same method for the respective datasets. On the basis of the obtained result, a new hybrid Attribute Evaluator method has been proposed which effectively integrates enhanced K-Means clustering with the CFS filter method and the BFS wrapper method. Furthermore, the hybrid method was evaluated with an improved decision tree classifier. The improved decision tree classifier combined clustering with classification. It was validated on 14 different chronic disease datasets and its performance was recorded. A very optimal and consistent classification performance was observed. The mean values for accuracy, specificity, sensitivity and f-score metrics were 96.7%, 96.5%, 95.6% and 96.2% respectively.

Cost-sensitive metaheuristic technique for credit card fraud detection

To prevent online financial losses, credit card frauds need to be addressed and it should be controlled. To minimize the credit card frauds, fraudulent credit card transactions should be identified and minimized. Keeping this fact in view, in this paper, a cost-sensitive metaheuristic technique (CSFPA: cost-sensitive learning flower pollination metaheuristic algorithm) is proposed to minimize the misclassification cost of credit card transactions from class imbalance data using the correlation-based feature section method (CFS), flower pollination algorithm (FPA), and cost-sensitive classifier. To identify the fraudulent transactions, the random forest (RF) ensemble classifier is used as a base learner in the cost-sensitive classifier for classification tasks and the proposed CSFPA technique has been tested on the Brazilian bank dataset. The performance of proposed CSFPA has been compared with cost insensitive techniques and cost-sensitive techniques such as NB (Naive Base), KNN (K-Nearest Neighbors), LB (Logit Boost), CSForest (Cost-sensitive decision forest), NBBFS (NB Best First Search) KNNGSS (KNN Greedy Stepwise Search), and CSFCS (CSForest Cuckoo Search). The experimental outcomes revealed that the proposed CSFPA technique has obtained encouraging outcomes for handling misclassification costs (Total cost and Avg. cost) and also has outperformed (FPR, Precision and Recall) compared to all the other approaches.

Simple Logistic Hybrid System Based on Greedy Stepwise Algorithm for Feature Analysis to Diagnose Parkinson’s Disease According to Gender

The incidence of Parkinson’s disease (PD) is higher in males than in females. This disease can be diagnosed based on gender through the automatic diagnostic system without visiting a specialist physician. For this purpose, the Simple Logistic hybrid system based on the greedy stepwise search algorithm (SLGS) is presented as a novel approach that involves feature analysis to diagnose PD by gender. Six feature groups were extracted from data of 252 subjects and were used to perform this analysis. When the SLGS hybrid system analysis begins, the effective features are first determined by the greedy stepwise search algorithm and diagnosis of PD was made according to the features that were determined using the Simple Logistic classifier. The performance results that were obtained from this system by gender were evaluated using many statistical metrics. The accuracy ratios were 88.71% and 87.15%, with the minimum number of features for males and females, respectively. Additionally, ROC and PRC area values for both genders approached the 0.9 band, showing that the class separation of patients and healthy individuals was nearly perfect. The performance results and perspective of this study were compared with the single published study that used the same data, and the SLGS hybrid system showed higher performance rates than the other study. In addition, the number of the subjects was higher in this study than in other studies and the SLGS system has not been used before in the literature in this field.

A Proposed Ensemble Model with Feature Selection Technique for Classification of Chronic Kidney Disease

Healthcare diagnosis system is very important and critical task in medical science for doctors and medical students. Chronic kidney disease is a very serious and dangerous problem which is directly related to the human life. In this research work, we have used data mining and feature selection technique to develop the robust and computationally efficient model for classifying chronic and non chronic kidney disease. An ensemble model is constructing through combination of two more similar types of trained model which helps to improve the performance. Feature selection is frequently used in machine learning area to raise a model with a few numbers of features which increase the performance of classification accuracy. The proposed feature selection techniques principle of Genetic Search (GS) and Greedy Stepwise Search (GSW). This proposed technique called GS-NB utilizes a pursuit methodology which is embedded in the Genetic Algorithm to select the features based on natural selection, the procedure that drives biological evolution. Then proposed technique called GSW-NB utilizes a search strategy that is included in the Greedy Stepwise to search the relevant feature based on problem solving heuristic for settling the locally ideal decision at each stage. The performance of suggested technique were estimated on Chronic Kidney Disease (CKD) classification problems and compared with proposed feature selection method. The classification techniques namely the Single Rule Classification (SRC), Conditional Inference Tree (CIT) and their ensemble model (SRC, CIT) have used for classification of CKD. The proposed ensemble model have used stacking learning technique which combines multiple classifiers, hence we can improve the performance of classifiers. The classifier performance is measured with observed accuracy, sensitivity and specificity. The experimental results demonstrated that the ensemble model (SRC, CIT) with GS-NB and GSW-NB can recognized CKD better than existing model. The proposed model can be beneficial and useful in medical science for identifying and diagnosis of chronic kidney disease.

A study of boosted evolutionary classifiers for detecting spam

PurposeEmail is a rapid and cheapest medium of sharing information, whereas unsolicited email (spam) is constant trouble in the email communication. The rapid growth of the spam creates a necessity to build a reliable and robust spam classifier. This paper aims to presents a study of evolutionary classifiers (genetic algorithm [GA] and genetic programming [GP]) without/with the help of an ensemble of classifiers method. In this research, the classifiers ensemble has been developed with adaptive boosting technique.Design/methodology/approachText mining methods are applied for classifying spam emails and legitimate emails. Two data sets (Enron and SpamAssassin) are taken to test the concerned classifiers. Initially, pre-processing is performed to extract the features/words from email files. Informative feature subset is selected from greedy stepwise feature subset search method. With the help of informative features, a comparative study is performed initially within the evolutionary classifiers and then with other popular machine learning classifiers (Bayesian, naive Bayes and support vector machine).FindingsThis study reveals the fact that evolutionary algorithms are promising in classification and prediction applications where genetic programing with adaptive boosting is turned out not only an accurate classifier but also a sensitive classifier. Results show that initially GA performs better than GP but after an ensemble of classifiers (a large number of iterations), GP overshoots GA with significantly higher accuracy. Amongst all classifiers, boosted GP turns out to be not only good regarding classification accuracy but also low false positive (FP) rates, which is considered to be the important criteria in email spam classification. Also, greedy stepwise feature search is found to be an effective method for feature selection in this application domain.Research limitations/implicationsThe research implication of this research consists of the reduction in cost incurred because of spam/unsolicited bulk email. Email is a fundamental necessity to share information within a number of units of the organizations to be competitive with the business rivals. In addition, it is continually a hurdle for internet service providers to provide the best emailing services to their customers. Although, the organizations and the internet service providers are continuously adopting novel spam filtering approaches to reduce the number of unwanted emails, the desired effect could not be significantly seen because of the cost of installation, customizable ability and the threat of misclassification of important emails. This research deals with all the issues and challenges faced by internet service providers and organizations.Practical implicationsIn this research, the proposed models have not only provided excellent performance accuracy, sensitivity with low FP rate, customizable capability but also worked on reducing the cost of spam. The same models may be used for other applications of text mining also such as sentiment analysis, blog mining, news mining or other text mining research.Originality/valueA comparison between GP and GAs has been shown with/without ensemble in spam classification application domain.

Malicious web pages detection using feature selection techniques and machine learning

In recent years, researchers have provided significant solutions to detect malicious web pages, still there are many open issues. This paper proposes a methodology for the effective detection of malicious web pages using feature selection methods and machine learning. Our methodology consists of three modules: feature selection, training and classification. To evaluate our methodology, six feature selection methods and eight supervised machine learning classifiers are used. Experiments are performed on the balanced binary dataset. It is found that by using feature selection methods, the classifiers achieved significant detection accuracy of 94-99% and above, error-rate of 0.19-5.55%, FPR of 0.006-0.094, FNR of 0.000-0.013 with minimum system overhead. Our multi-model system using majority voting classifier and wrapper+Naive Bayes feature selection method with GreedyStepwise search technique using only 15 features achieved a highest accuracy of 99.15%, FPR of 0.017 and FNR of 0.000.

Prediction of sheep carcass traits from early-life records using machine learning

Currently hot carcass weight (HCW) and fat score jointly indicate the price grid for sheep meat in Australia. However, experts in the field believe that soon, yield and quality traits such as intramuscular fat (IMF), greville rule fat depth (GRFAT), computed tomography lean meat yield (CTLEAN), and loin weight (LW) are likely to play a role in pricing. Having an accurate prediction of these traits earlier in the life of an animal will allow sheep producers to adjust their management practices in order to achieve the target market requirements. Management, genetics, pasture and climate factors, influence these traits directly and epistatically. Traditional prediction methods may not be powerful enough to capture complex interactions while avoiding overfitting. In this case, learning algorithms that can learn from the current data to predict the animal’s future performance offers promise. In this study, five different types of Machine Learning (ML) algorithm, namely Deep Learning (DL), Gradient Boosting Tree (GBT), K-Nearest Neighbour (KNN), Model Tree (MT), and Random Forest (RF) were employed to predict HCW, IMF, GRFAT, LW and CTLEAN and their performances were compared against linear regression (LR) as the gold standard of multinomial prediction. Four scenarios representing different numbers of weight recordings -from a total of 9 weight measures taken between birth (WT1) and pre-slaughter (WT9)- were used to inform the algorithms and all models were trained and tested under equal conditions with identical training and testing sets. Selection of the most effective subset of predictor features were completed via greedy stepwise search among all the available features jointly with expert opinion. In predicting all the traits, RF was superior while LR and KNN showed the lowest prediction performance. When using the final model for predicting on an independent test set, the scenario with the most accurate prediction performance differed across traits. IMF and GRFAT were most accurately predicted when using birth, weaning, and pre-slaughter weights, while the most accurate scenario for HCW, LW and CTLEAN utilised weaning, six monthly weight measures after weaning and pre-slaughter weight. Across all scenarios the least accurate prediction was for IMF.

Analysing user sentiment of Indian movie reviews

Purpose To be sustainable and competitive in the current business environment, it is useful to understand users’ sentiment towards products and services. This critical task can be achieved via natural language processing and machine learning classifiers. This paper aims to propose a novel probabilistic committee selection classifier (PCC) to analyse and classify the sentiment polarities of movie reviews. Design/methodology/approach An Indian movie review corpus is assembled for this study. Another publicly available movie review polarity corpus is also involved with regard to validating the results. The greedy stepwise search method is used to extract the features/words of the reviews. The performance of the proposed classifier is measured using different metrics, such as F-measure, false positive rate, receiver operating characteristic (ROC) curve and training time. Further, the proposed classifier is compared with other popular machine-learning classifiers, such as Bayesian, Naïve Bayes, Decision Tree (J48), Support Vector Machine and Random Forest. Findings The results of this study show that the proposed classifier is good at predicting the positive or negative polarity of movie reviews. Its performance accuracy and the value of the ROC curve of the PCC is found to be the most suitable of all other classifiers tested in this study. This classifier is also found to be efficient at identifying positive sentiments of reviews, where it gives low false positive rates for both the Indian Movie Review and Review Polarity corpora used in this study. The training time of the proposed classifier is found to be slightly higher than that of Bayesian, Naïve Bayes and J48. Research limitations/implications Only movie review sentiments written in English are considered. In addition, the proposed committee selection classifier is prepared only using the committee of probabilistic classifiers; however, other classifier committees can also be built, tested and compared with the present experiment scenario. Practical implications In this paper, a novel probabilistic approach is proposed and used for classifying movie reviews, and is found to be highly effective in comparison with other state-of-the-art classifiers. This classifier may be tested for different applications and may provide new insights for developers and researchers. Social implications The proposed PCC may be used to classify different product reviews, and hence may be beneficial to organizations to justify users’ reviews about specific products or services. By using authentic positive and negative sentiments of users, the credibility of the specific product, service or event may be enhanced. PCC may also be applied to other applications, such as spam detection, blog mining, news mining and various other data-mining applications. Originality/value The constructed PCC is novel and was tested on Indian movie review data.

A modified content-based evolutionary approach to identify unsolicited emails

This computational research seeks to classify unsolicited versus legitimate emails. A modified version of an existing genetic programming (GP) classifier—i.e., modified genetic programming (MGP)—is implemented to build an ensemble of classifiers to identify unsolicited emails. The proposed classifier is assessed using informative features extracted from two corpora (Enron and SpamAssassin) with the help of the greedy stepwise feature search method. Further, a comparative study is performed with other popular classifiers, such as Bayesian network, naive Bayes, decision tree, random forest (RF), support vector machine (SVM), and GP. Further the results are validated with 20-fold cross-validation and paired T test. The results prove that the proposed classifier performs better in terms of accuracy and false-positive detection in comparison with the other machine learning classifiers tested in this study. Using different training and testing a set of email files from the Enron corpus, ensemble-based classifiers, such as boosted SVM, boosted Bayesian, boosted naive Bayesian, RF, and the proposed MGP classifier, are tested and compared on all metrics, including training and testing time. The findings suggest that the MGP classifier with the greedy stepwise feature search method offers an improvement over alternative methods in detecting unsolicited emails.

Input variable selection with greedy stepwise search algorithm for analysing the probability of fish occurrence: A case study for Alburnoides mossulensis in the Gamasiab River, Iran

Input variable selection using data-driven combined with statistical methods have received more attention to analyse the probability of freshwater organisms’ occurrence. Eight different sampling sites (from the source to the mouth of Gamasiab River basin, Iran) were considered to study the occurrence of Alburnoides mossulensis during one year study period (2008–2009). A set of river characteristics together with abundance of target fish (based on presence/absence data) were recorded at each sampling site. Logistic regression was optimized with an input variable selection, greedy stepwise search algorithm, to select the most important explanatory variables for analysing the occurrence of fish. According to the optimization method, almost one-third of total recorded variables in the sampling sites including electric conductivity (EC), bicarbonate (HCO3–), river width, river depth, water temperature, pH, sulphate (SO42−) and orthophosphate (PO43−-P) might influence the probability of occurrence of fish in the river while based on the outcomes of binary logistic regression model, electrical conductivity and bicarbonate were the most important ones (p < 0.05 for both variables).

The Impact of Feature Selection on Urban Land Cover Classification

Many of the studies in the literature about land cover classification are focused on the feature extraction and classification rather than feature selection. In this paper, the impact of feature selection on urban land cover classification is extensively analyzed. Three types of features namely spectral, texture, and size/shape features are used for this analysis. This analysis is carried out using three variations of a filter based feature selection method and three widely-known classification algorithms. The feature selection method used for the comparison is a multivariate filter method namely correlation-based feature subset selection where a feature subset evaluator and a search method are integrated. Best first search, genetic search, and greedy stepwise search are three different search methods used for this integration. The classification algorithms employed are Bayesian network, random forest, and support vector machine. The experimental results explicitly indicate that feature selection improves classification accuracy in all cases. Besides, according to the experimental results, random forest classifier is the most successful one among these three classifiers while both feature selection is applied and not applied. Largest improvement in the classification performance is obtained when greedy stepwise search based feature selection method and support vector machine classifier is applied together. Also, the contribution of spectral features to the performance of classification is more than size/shape and texture features.

Diagnosis of Chronic Kidney Disease Based on Support Vector Machine by Feature Selection Methods.

As Chronic Kidney Disease progresses slowly, early detection and effective treatment are the only cure to reduce the mortality rate. Machine learning techniques are gaining significance in medical diagnosis because of their classification ability with high accuracy rates. The accuracy of classification algorithms depend on the use of correct feature selection algorithms to reduce the dimension of datasets. In this study, Support Vector Machine classification algorithm was used to diagnose Chronic Kidney Disease. To diagnose the Chronic Kidney Disease, two essential types of feature selection methods namely, wrapper and filter approaches were chosen to reduce the dimension of Chronic Kidney Disease dataset. In wrapper approach, classifier subset evaluator with greedy stepwise search engine and wrapper subset evaluator with the Best First search engine were used. In filter approach, correlation feature selection subset evaluator with greedy stepwise search engine and filtered subset evaluator with the Best First search engine were used. The results showed that the Support Vector Machine classifier by using filtered subset evaluator with the Best First search engine feature selection method has higher accuracy rate (98.5%) in the diagnosis of Chronic Kidney Disease compared to other selected methods.

A Meta-Model Implementation with Tabu Search Technique to Determine the Buying Pattern of Online Customers

Customers shopping pattern depends on features like demography, psychology, culture etc. Observing factors that affect online shopping behavior helps in classifying customers and devising proper marketing strategy. In this paper a metamodel was proposed to classify Indian online shoppers based on their online shopping pattern. Here 163 observations with 14 attributes related to online shopping data were gathered through a questionnaire. Online shopping overhead was relatively reduced by removing irrelevant features from the dataset by using Greedy Stepwise search and Tabu search. This reduced set of shopping data is applied to RBF classifier to correctly predict customer shopping trends. Results were evaluated using some vital performance metrics to determine classification efficiency. The proposed model using Tabu Search was inferred to be more performance driven since the classification accuracy with Tabu search was 88.35% while the error rate was reduced to 0.0786. The model development time was found to be just 1.23 sec with Tabu search. The Mathews Coefficient was determined to be 0.774 with Tabu search. The proposed work can also be extended to classify other real time data sets. Classification accuracy may further be enhanced by adding physiographic and cultural inputs to the proposed classifier.

Nature Inspired Feature Selection Approach for Effective Intrusion Detection

Objectives: To reduce the dimensionality of network traffic dataset by selecting the relevant and irredundant features for accurate and quick intrusion detection. To achieve the target, we proposed a new feature selection approach based on the nature. Methods/Statistical Analysis: The proposed Modified CuttleFish Algorithm (MCFA) approach plays a crucial role in intrusion detection by selecting appropriate subset of most relevant features from huge amount of dataset. Griewank fitness function is used to calculate the fitness of the modified cuttlefish algorithm. Naive bayes classifier is employed at the generated subset of features from benchmark KDD 99 dataset in WEKA data mining tool Compare the results of proposed approach with the existing approaches of WEKA and Improved Cuttlefish Algorithm (ICFA), with different performance metrics. Findings: As per the outcomes are obtained by the WEKA Experimenter with the 9 feature selection approaches on KDD 99 10% training dataset, it has been observed that the Consistency Subset feature selection approach with Greedy stepwise search method gives higher accurate results than other approaches and from the literature survey has been found that the ICFA performs well, but still there is problem of low True positive (TP) rate and False negative (FN) rate. These problems are addressed by the proposed feature selection approach which outer perform best from others in accuracy rate (91.79%), True positive rate (0.947), false positive rate (0.025) and ROC area (0.9982) with the minimum amount of time at 19 relevant subset of features instead 41 features. As the consequence, the proposed approach is novel from existing approaches to increase the intrusion detection rate and discard the redundant and irrelevant subset of features. Application/Improvements: MCFA has improved intrusion detection rate by increasing the TP rate and decreasing the FP rate, so MCFA can be used for the real time applications of intrusion detection system. Keywords: Accuracy, Dimensionality, Feature Selection Approach, FP Rate, Intrusion Detection, Modified Cuttlefish Algorithm, TP Rate

Interaction between feature subset selection techniques and machine learning classifiers for detecting unsolicited emails

Detection of the spam emails within a set of email files has become challenging task for researchers. Identification of an effective classifier is based not only on high accuracy of detection but also on low false alarm rates, and the need to use as few features as possible. In view of these challenges, this research examines the effects of using features selected by four feature subset selection methods (i.e. Genetic, Greedy Stepwise, Best First, and Rank Search) on popular Machine Learning Classifiers like Bayesian, Naive Bayes, Support Vector Machine, Genetic Algorithm, J48 and Random Forest. Tests were performed on three different publicly available spam email datasets: "Enron", "SpamAssassin" and "LingSpam". Results show that, Greedy Stepwise Search method is a good method for feature subset selection for spam email detection. Among the Machine Learning Classifiers, Support Vector Machine has been found to be the best classifier both in terms of accuracy and False Positive rate. However, results of Random Forest were very close to that of Support Vector Machine. The Genetic classifier was identified as a weak classifier.

Development of a 5 year life expectancy index in older adults using predictive mining of electronic health record data

Incorporating accurate life expectancy predictions into clinical decision making could improve quality and decrease costs, but few providers do this. We sought to use predictive data mining and high dimensional analytics of electronic health record (EHR) data to develop a highly accurate and clinically actionable 5 year life expectancy index. We developed the index using EHR data for 7463 patients ≥50 years old with ≥1 visit(s) in 2003 to a large, academic, multispecialty group practice. We extracted 980 attributes from the EHRs of the practices and affiliated hospitals. Correlation feature selection with greedy stepwise search was used to find the attribute subset with best average merit. Rotation forest ensembling with alternating decision tree as underlying classifier was used to predict 5 year mortality. Model performance was compared with the modified Charlson Comorbidity Index and the Walter life expectancy method. Within 5 years of the last visit in 2003, 838 (11%) patients had died. The final model included 24 attributes: two demographic (age, sex), 10 comorbidity (eg, cardiovascular disease), one vital sign (mean diastolic blood pressure), two medications (loop diuretic use, digoxin use), six laboratory (eg, mean albumin), and three healthcare utilization (eg, the number of hospitalizations 1 year prior to the last visit in 2003). The index showed very good discrimination (c-statistic 0.86) and outperformed comparators. The EHR based index successfully distinguished adults ≥50 years old with life expectancy >5 years from those with life expectancy ≤5 years. This information could be used clinically to optimize preventive service use (eg, cancer screening in the elderly).

Breast Cancer Prediction System using Feature Selection and Data Mining Methods

Cancer is the second most common cause of death worldwide with an estimated 7.9 million of deaths in 2007. This number is projected to rise further and reach 12 million deaths in 2030 which makes cancer a major public health issue. Early diagnosis and more recent tools of managing cancer have shown to significantly improve the chances of survival and have brought new hope to patients. The identity of cancer from various factors or symptoms is a multifaceted issue which is not free from false presumptions often accompanied by volatile effects. In this paper, we have proposed three different feature selection method rank search, genetic search and greedy step wise search methods to identify the potential attributes from the Breast Cancer dataset using the classification of heart attack using data mining techniques. This breast cancer databases was obtained from the University of Wisconsin Hospitals, Madison from Dr. William H. Wolberg. Attributes 2 through 10 have been used to represent instances. Each instance has one of 2 possible classes: benign or malignant. Number of instances: 699 we have investigated six different classification data mining techniques such as BayesNet, AttributeSelectedClassifier, J48, ClassificationviaRegression, Logistic, and OneR. The result shows that the three different set of potential attributes are obtained through rank search, genetic search and greedy step. It is observed that the performance and the time taken by each classification algorithms are significantly improved after feature selection and the bayesnet classifier outperforms the remaining algorithms used in this paper. Keywords: Breast Cancer, Data Mining, Classification, BayesNet, ClassificationviaRegression, Logistic, Rank Search, Genetic Search , Greedy Stepwise Search

Comparison of modelling techniques to predict macroinvertebrate community composition in rivers of Ethiopia

In order to fulfil the millennium development goals and to ensure environmental sustainability in Ethiopia, ecological indicator systems can support river managers to analyse the status of watercourses and to select critical restoration actions. In order to use macroinvertebrates as river water quality monitoring and assessment tools, Ethiopia needs data from reference as well as disturbed conditions of surface water ecosystems. Macroinvertebrates, structural and physical–chemical data were in this context collected in the Gilgel Gibe river basin in South-Western Ethiopia during the period 2005–2008. In the next stage, ecological metrics were compared for their assessment relevance. In the present paper, classification trees and support vector machines were used to induce models describing the relation between the river characteristics and the ecological conditions of these streams. Greedy stepwise and genetic search algorithms improved the performance and easy interpretation of these models by making a selection of the variables that were used as input of these models. The developed models allowed to identify the major variables affecting river quality. These tools can support river managers in their decision-making regarding the status of rivers and potential restoration options, for example by providing rules concerning critical values of major river characteristics at which certain actions should be undertaken.