The Method of Attribute Reduction Based on Decision Tree

Distance metric learning-based multi-granularity neighborhood rough sets for attribute reduction

Background: Assessment of liver fibrosis is a vital need for enabling therapeutic decisions and prognostic evaluations of chronic hepatitis. Liver biopsy is considered the definitive investigation for assessing the stage of liver fibrosis but it carries several limitations. FIB-4 and APRI also have a limited accuracy. The National Committee for Control of Viral Hepatitis (NCCVH) in Egypt has supplied a valuable pool of electronic patients’ data that data mining techniques can analyze to disclose hidden patterns, trends leading to the evolution of predictive algorithms. Aim: to collaborate with physicians to develop a novel reliable, easy to comprehend noninvasive model to predict the stage of liver fibrosis utilizing routine workup, without imposing extra costs for additional examinations especially in areas with limited resources like Egypt. Methods: This multi-centered retrospective study included baseline demographic, laboratory, and histopathological data of 69106 patients with chronic hepatitis C. We started by data collection preprocessing, cleansing and formatting for knowledge discovery of useful information from Electronic Health Records EHRs. Data mining has been used to build a decision tree (Reduced Error Pruning tree (REP tree)) with 10-fold internal cross-validation. Histopathology results were used to assess accuracy for fibrosis stages. Machine learning feature selection and reduction (CfsSubseteval / best first) reduced the initial number of input features (N=15) to the most relevant ones (N=6) for developing the prediction model. Results: In this study, 32419 patients had F(0-1), 25073 had F(2) and 11615 had F(3-4). FIB-4 and APRI revalidation in our study showed low accuracy and high discordance with biopsy results, with overall AUC 0.68 and 0.58 respectively. Out of 15 attributes machine learning selected Age, AFP, AST, glucose, albumin, and platelet as the most relevant attributes. Results for REP tree indicated an overall classification accuracy up to 70% and ROC Area 0.74 which was not nearly affected by attribute reduction, and pruning . However attribute reduction, and tree pruning were associated with simpler model easy to understand by physician with less time for execution. Conclusion: This study we had the chance to study a large cohort of 69106 chronic hepatitis patients with available liver biopsy results to revise and validate the accuracy of FIB-4 and APRI. This study represents the collaboration between computer scientist and hepatologists to provide clinicians with an accurate novel and reliable, noninvasive model to predict the stage of liver fibrosis.

In traditional decision (classification) tree algorithms, the label is assumed to be a categorical (class) variable. When the label is a continuous variable in the data, two possible approaches based on existing decision tree algorithms can be used to handle the situations. The first uses a data discretization method in the preprocessing stage to convert the continuous label into a class label defined by a finite set of nonoverlapping intervals and then applies a decision tree algorithm. The second simply applies a regression tree algorithm, using the continuous label directly. These approaches have their own drawbacks. We propose an algorithm that dynamically discretizes the continuous label at each node during the tree induction process. Extensive experiments show that the proposed method outperforms the preprocessing approach, the regression tree approach, and several nontree-based algorithms.

Recent advances in computing, communications, digital storage technologies, and high-throughput data-acquisition technologies, make it possible to gather and store incredible volumes of data. It creates unprecedented opportunities for large-scale knowledge discovery from huge database. Data mining (DM) technology has emerged as a means of performing this discovery. There are countless researchers working on designing efficient data mining techniques, methods, and algorithms. Many data mining methods and algorithms have been developed and applied in a lot of application fields [1]. Unfortunately, most data mining researchers pay much attention to technique problems for developing data mining models and methods, while little to basic issues of data mining [2]. In this talk, some basic issues of data mining are addressed. What is data mining? What is the product of a data mining process? What are we doing in a data mining process? What is the rule we should obey in a data mining process? Through analyzing existing data mining methods, and domain-driven (or user-driven) data mining models [3-5], we find that we should take a data mining process as a process of knowledge transformation. Based on this understanding of data mining, a conceptual data mining model of domain-oriented data-driven data mining (3DM) is proposed [2]. The relationship between traditional domain-driven (or user-driven) data mining models and our proposed 3DM model is also analyzed. Some domain-oriented data-driven data mining algorithms for mining such knowledge as default rule [6], decision tree [7], and concept lattice [8] from database are proposed. The experiment results for these algorithms are also shown to illustrate the efficiency and performance of the knowledge acquired by our 3DM data mining algorithms.

Aim: Chronic kidney disease (CKD) is among the main 20 reasons for death worldwide and influences around 10% of the world grown-up populace. CKD is an issue that upsets typical kidney work. The main objective of this study aims to find the best-suited algorithm that will give us the most ideal prediction. We will be comparing Novel Decision Tree with Linear Regression to find out which of these can give us the best accuracy. Material and Methods: The study used 322 samples with Novel Decision Tree and Linear Regression is executed with varying training and testing splits for predicting the accuracy for kidney disease prediction with the G-power value of 80% and the kidney datasets were collected from various web sources with recent study findings and threshold 0.05%, confidence interval 95% mean and standard deviation. The performance of the classifiers are evaluated based on their accuracy rate using the chronic kidney disease dataset. Results: The accuracy of predicting kidney disease in Novel Decision Tree (96.66%) and Linear Regression (85.25%) is obtained. There is a statistical 2-tailed significant difference in accuracy for two algorithms is 0.000 (p<0.05) by performing independent samples t-tests. Conclusion: This study concludes that the Prediction of Kidney disease using the Novel Decision Tree (DT) algorithm appears to be significantly better than the Linear Regression(LR) with improved accuracy.

Results in heart disease classification that are inaccurate and have low accuracy can endanger the patient's life. Some parameters in the algorithm model also influence classification. This study compares the Decision Tree and Random Forest algorithms for heart disease. The influence of maximum depth on heart disease classification also has significant implications. If the maximum depth is not set correctly, the classification results can be inaccurate and lead to incorrect diagnoses. This study uses five data split schemes, namely 60%: 40%, 70%: 30%, 75%: 25%, 80%: 20%, 90%: 10% and tested with different max depth parameters, namely max depth = 3, 4, 5, 6, and 7. This research produces the best accuracy using the 90%:10% scheme and max depth = 7 with the best accuracy result using the Random Forest algorithm of 99.29% while the Decision Tree algorithm is 98.05%. Then the precision and recall value of the Random Forest algorithm is 99% while the Decision Tree is 98%. The results of computation time using Decision Tree are faster than using Random Forest with a computation time for training data of 0.0075 s, while the testing data are 0.009 s. In future research, research can be conducted on the effect of other parameters by testing using several data sets.

An improved attribute reduction scheme with covering based rough sets

A scalable decision tree system and its application in pattern recognition and intrusion detection

Manufacturing industries facing problem in optimal selection of process parameters in machining process. Finding optimum process parameters for achieving maximum Material Removal Rate and minimum Surface Roughness is a challenging task and it requires lot of time and energy for experimentation trails or experience. It wastes lot of resources and money, sometimes ends up with negative results. To overcome the above issue, this paper presents an algorithm for prediction of Surface Roughness and Material Removal rate using Decision Tree Algorithm and Naive Bayes Algorithm without experimentation. Lot of resources and time can be saved using these machine learning algorithms. In this paper, Material removal rate and Surface roughness of EDM machining of Aluminum composites is predicted using Decision tree algorithm and Naive Bayes algorithm. Then the model can be used to predict the Material Removal Rate and Surface finish of any combination process parameters before machining process.

Attribute reduction plays an important role in knowledge discovery and data mining. Confronted with data characterized by the interval and missing values in many data analysis tasks, it is interesting to research the attribute reduction for interval-valued data with missing values. Uncertainty measures can supply efficient viewpoints, which help us to disclose the substantive characteristics of such data. Therefore, this paper addresses the attribute reduction problem based on uncertainty measure for interval-valued data with missing values. At first, an uncertainty measure is provided for measuring candidate attributes, and then an efficient attribute reduction algorithm is developed for the interval-valued data with missing values. To improve the efficiency of attribute reduction, the objects that fall within the positive region are deleted from the whole object set in the process of selecting attributes. Finally, experimental results demonstrate that the proposed algorithm can find a subset of attributes in much shorter time than existing attribute reduction algorithms without losing the classification performance.

Mixture design optimization and machine learning-based prediction of Al-Mg alloy composite reinforced by Zn nanoparticles: A molecular dynamics study

An incremental approach to attribute reduction from dynamic incomplete decision systems in rough set theory

A systematic study on attribute reduction with rough sets based on general binary relations

The virtualization of computing resources provided by containers has gained increasing attention and has been widely used in cloud computing. This new demand for container technology has been growing and the use of Docker and Kubernetes is considerable. According to recent technology surveys, containers are now mainstream. However, currently, one of the major challenges rises from the fact that multiple containers, with different owners, may cohabit on the same host. In container-based multi-tenant environments, security issues are of major concern. In this paper we investigate the performance of container-level anomaly-based intrusion detection systems for multi-tenant applications. We investigate the use of Bag of System Calls (BoSC) technique and the sliding window with the classifier and we consider eight machine learning algorithms for classification purposes. We show that among the eight machine learning algorithms, the best classification results are obtained with Decision Tree and Random Forest which lead to an F-Measure of 99.8%, using a sliding window with a size of 30 and the BoSC algorithm in both cases. We also show that, although both Decision Tree and Random Forest algorithms leads to the best classification results, the Decision Tree algorithm has a shorter execution time and consumes less CPU and memory than the Random Forest.

The study objective is to validate a mathematical model for prognosis of progression of the acute period of myocardial infarction. Materials and methods . We examined 135 patients with acute Q-wave myocardial infarction of the left ventricle aged between 34 and 88 years (mean age 66, standard deviation 12 years), among them 64 women and 71 men. For prognosis of the outcome of the acute period of myocardial infarction we used an approach based on intellectual data analysis (data mining) in combination with mathematical methods based on decision trees. Results. Using decision tree algorithms, we singled out laboratory parameters (attributes) which were subsequently used as input. Adequacy of classification of these attributes was determined by a contingency table. Accuracy of the obtained calculation results was 95.56 % demonstrating good agreement between the model and observed data. In a decision tree visualization, the most significant 8 laboratory parameters were determined. Significance of NO 2 metabolite was 24.9 %, triglycerides – 16.7 %, urea – 14.8 %, erythrocytes – 11.2 %, alanine aminotransferase – 9.4 %, very low density lipoproteins – 9.4 %, creatinine – 8.5 %, prothrombin index – 5.1 %. In the Rules tab, only rules 4 and 9 can be used with confidence, because their confidence level approaches 100 %, and effect cost for the fact of death was 33.59 % and 32.03 %, respectively. Conclusion. Using a decision tree algorithm, we determined prognostically significant factors for progression of acute myocardial infarction. The following set of parameters predicts unfavorable outcome (death) with 95.56 % accuracy: NO 2 < level 22.755 mmol/l, triglycerides ≥ 1.565 mmol/l, erythrocytes < 4.91 M/uL, alanine aminotransferase < 1.23 mmol/l, urea < 7.05 mmol/l, very low-density lipoproteins < 0.965 mmol/l, creatinine ≥ 91.55 µmol/l, NO 2 level ≥ 22.755 mmol/l predicts a favorable outcome with 95.56 % accuracy.