Rule-based Learning Algorithm Research Articles

Classification of Non-Functional Requirements From IoT Oriented Healthcare Requirement Document.

Internet of Things (IoT) involves a set of devices that aids in achieving a smart environment. Healthcare systems, which are IoT-oriented, provide monitoring services of patients' data and help take immediate steps in an emergency. Currently, machine learning-based techniques are adopted to ensure security and other non-functional requirements in smart health care systems. However, no attention is given to classifying the non-functional requirements from requirement documents. The manual process of classifying the non-functional requirements from documents is erroneous and laborious. Missing non-functional requirements in the Requirement Engineering (RE) phase results in IoT oriented healthcare system with compromised security and performance. In this research, an experiment is performed where non-functional requirements are classified from the IoT-oriented healthcare system's requirement document. The machine learning algorithms considered for classification are Logistic Regression (LR), Support Vector Machine (SVM), Multinomial Naive Bayes (MNB), K-Nearest Neighbors (KNN), ensemble, Random Forest (RF), and hybrid KNN rule-based machine learning (ML) algorithms. The results show that our novel hybrid KNN rule-based machine learning algorithm outperforms others by showing an average classification accuracy of 75.9% in classifying non-functional requirements from IoT-oriented healthcare requirement documents. This research is not only novel in its concept of using a machine learning approach for classification of non-functional requirements from IoT-oriented healthcare system requirement documents, but it also proposes a novel hybrid KNN-rule based machine learning algorithm for classification with better accuracy. A new dataset is also created for classification purposes, comprising requirements related to IoT-oriented healthcare systems. However, since this dataset is small and consists of only 104 requirements, this might affect the generalizability of the results of this research.

Dealing with data intrinsic difficulties by learning an interPretable Ensemble Rule Learning (PERL) model

One of the challenges of learning algorithms is data imbalance, especially data intrinsic difficulties such as borderline, rare, and outlier examples. In order to improve the performance of learning algorithms in such cases, various solutions have been proposed, one of which is ensemble learning. Despite the good results of ensemble learning, their results are not interpretable. In some domains, such as medicine, banking, and telecommunications, the interpretability of the final model results is very important. For this purpose, rule-based learning algorithms and decision trees are two appropriate approaches. Therefore, an interPretable Ensemble Rule Learning (PERL) model is proposed in this paper. PERL has three main phases: sampling, rule learning, and rule processing. We reduce the negative effects of data intrinsic difficulties in each phase by an appropriate solution. PERL is compared with several interpretable ensemble models on ten imbalanced real-world datasets. According to the obtained results, PERL outperforms significantly the other models based on the f1-score measure.

Interpretable genotype-to-phenotype classifiers with performance guarantees

Understanding the relationship between the genome of a cell and its phenotype is a central problem in precision medicine. Nonetheless, genotype-to-phenotype prediction comes with great challenges for machine learning algorithms that limit their use in this setting. The high dimensionality of the data tends to hinder generalization and challenges the scalability of most learning algorithms. Additionally, most algorithms produce models that are complex and difficult to interpret. We alleviate these limitations by proposing strong performance guarantees, based on sample compression theory, for rule-based learning algorithms that produce highly interpretable models. We show that these guarantees can be leveraged to accelerate learning and improve model interpretability. Our approach is validated through an application to the genomic prediction of antimicrobial resistance, an important public health concern. Highly accurate models were obtained for 12 species and 56 antibiotics, and their interpretation revealed known resistance mechanisms, as well as some potentially new ones. An open-source disk-based implementation that is both memory and computationally efficient is provided with this work. The implementation is turnkey, requires no prior knowledge of machine learning, and is complemented by comprehensive tutorials.

Graph- and rule-based learning algorithms: a comprehensive review of their applications for cancer type classification and prognosis using genomic data.

Cancer is well recognized as a complex disease with dysregulated molecular networks or modules. Graph- and rule-based analytics have been applied extensively for cancer classification as well as prognosis using large genomic and other data over the past decade. This article provides a comprehensive review of various graph- and rule-based machine learning algorithms that have been applied to numerous genomics data to determine the cancer-specific gene modules, identify gene signature-based classifiers and carry out other related objectives of potential therapeutic value. This review focuses mainly on the methodological design and features of these algorithms to facilitate the application of these graph- and rule-based analytical approaches for cancer classification and prognosis. Based on the type of data integration, we divided all the algorithms into three categories: model-based integration, pre-processing integration and post-processing integration. Each category is further divided into four sub-categories (supervised, unsupervised, semi-supervised and survival-driven learning analyses) based on learning style. Therefore, a total of 11 categories of methods are summarized with their inputs, objectives and description, advantages and potential limitations. Next, we briefly demonstrate well-known and most recently developed algorithms for each sub-category along with salient information, such as data profiles, statistical or feature selection methods and outputs. Finally, we summarize the appropriate use and efficiency of all categories of graph- and rule mining-based learning methods when input data and specific objective are given. This review aims to help readers to select and use the appropriate algorithms for cancer classification and prognosis study.

Automated detection of unusual soil moisture probe response patterns with association rule learning

In-situ field monitoring networks generate vast quantities of continuous data can help to improve the design, management, operation and maintenance of Green Infrastructure (GI) systems. However, such actions require efficient and reliable quality assurance quality control (QAQC). In this paper, we develop a rule-based learning algorithm involving Dynamic Time Warping (DTW) to investigate the feasibility of detecting anomalous responses from soil moisture probes using data collected from a GI site in Milwaukee, WI. As an enhancement to traditional QAQC methods which rely on individual time steps, this method converts the continuous time series into event sequences from which response patterns can be detected. Association rules are developed on both environmental features and event features. The results suggest that this method could be used to identify abnormal change patterns as compared to intra-site historical observations. Though developed for soil moisture, this method could easily be extended to apply on other continuous environmental datasets.

Comparison of rule induction, decision trees and formal concept analysis approaches for classification

Rule-based learning algorithms have higher transparency and easiness to interpret in comparison with neural networks and deep learning algorithms. These properties make it possible to effectively use such algorithms to solve descriptive tasks of data mining. The choice of an algorithm depends also on its ability to solve predictive tasks. The article compares the quality of the solution of the problems with binary and multiclass classification based on the experiments with six datasets from the UCI Machine Learning Repository. The authors investigate three algorithms: Ripper (rule induction), C4.5 (decision trees), In-Close (formal concept analysis). The results of the experiments show that In-Close demonstrates the best quality of classification in comparison with Ripper and C4.5, however the latter two generate more compact rule sets.

Validation of natural language processing (NLP) for automated ascertainment of EGFR and ALK tests in SEER cases of non-small cell lung cancer (NSCLC).

6528 Background: The Surveillance, Epidemiology, and End Results (SEER) registries lack information on the Epidermal Growth Factor Receptor (EGFR) mutation and Anaplastic Lymphoma Kinase (ALK) gene rearrangement test results. With the goal of enabling population-based outcomes research in molecularly selected NSCLC subgroups, we conducted a validation study of NLP for ascertainment of EGFR and ALK testing from electronic pathology reports (e-paths) of patients included in the Seattle-Puget Sound (SPS) and Kentucky Cancer (KCR) SEER registries. Methods: We obtained 4,278 and 1,041 e-paths pertaining to 1,634 and 565 patients with stage IV non-squamous NSCLC diagnosed from 1/1/2011 to 12/31/2013 and included in the SPS and KCR registries, respectively. Two oncologists independently reviewed all reports to generate a gold-standard dataset. We used 855 of the SPS reports to train hybrid rule-based and machine learning algorithms for detection of test status (reported vs. not reported), and test result if reported (positive vs negative) for EGFR mutational analysis and ALK testing by FISH, IHC, or gene sequencing. In the remaining 3,423 SPS reports, we conducted a 5-fold cross-validation analysis to estimate the internal NLP sensitivity, specificity, positive predictive value, and negative predictive value for test status and results, respectively. We used a hierarchical rules system to assess the NLP accuracy at the patient level. For external validation, we repeated all analyses in the KCR dataset. Results: In the SPS internal validation report sample, the validity metrics ranged from 97% to 99% for EGFR and ALK test status, and from 95% to 100% for EGFR and ALK test results, respectively. In the KCR external validation report sample, the metrics ranged from 74% to 96% for EGFR and ALK test status, and 2% to 100% for test results, respectively. At the patient level, the NLP accuracy for EGFR and ALK was 95% and 96% (SPS cohort), and 70% and 72% (KCR cohort) respectively. Conclusions: NLP is a valid method for determining EGFR and ALK test status and results for patients included in SEER registries with access to e-path, but the algorithms likely need to be registry-specific.

교통사고 데이터의 마이닝을 위한 연관규칙 학습기법과 서브그룹 발견기법의 비교

Traffic accident is one of the major cause of death worldwide for the last several decades. According to the statistics of world health organization, approximately 1.24 million deaths occurred on the world's roads in 2010. In order to reduce future traffic accident, multipronged approaches have been adopted including traffic regulations, injury-reducing technologies, driving training program and so on. Records on traffic accidents are generated and maintained for this purpose. To make these records meaningful and effective, it is necessary to analyze relationship between traffic accident and related factors including vehicle design, road design, weather, driver behavior etc. Insight derived from these analysis can be used for accident prevention approaches. Traffic accident data mining is an activity to find useful knowledges about such relationship that is not well-known and user may interested in it. Many studies about mining accident data have been reported over the past two decades. Most of studies mainly focused on predict risk of accident using accident related factors. Supervised learning methods like decision tree, logistic regression, k-nearest neighbor, neural network are used for these prediction. However, derived prediction model from these algorithms are too complex to understand for human itself because the main purpose of these algorithms are prediction, not explanation of the data. Some of studies use unsupervised clustering algorithm to dividing the data into several groups, but derived group itself is still not easy to understand for human, so it is necessary to do some additional analytic works. Rule based learning methods are adequate when we want to derive comprehensive form of knowledge about the target domain. It derives a set of if-then rules that represent relationship between the target feature with other features. Rules are fairly easy for human to understand its meaning therefore it can help provide insight and comprehensible results for human. Association rule learning methods and subgroup discovery methods are representing rule based learning methods for descriptive task. These two algorithms have been used in a wide range of area from transaction analysis, accident data analysis, detection of statistically significant patient risk groups, discovering key person in social communities and so on. We use both the association rule learning method and the subgroup discovery method to discover useful patterns from a traffic accident dataset consisting of many features including profile of driver, location of accident, types of accident, information of vehicle, violation of regulation and so on. The association rule learning method, which is one of the unsupervised learning methods, searches for frequent item sets from the data and translates them into rules. In contrast, the subgroup discovery method is a kind of supervised learning method that discovers rules of user specified concepts satisfying certain degree of generality and unusualness. Depending on what aspect of the data we are focusing our attention to, we may combine different multiple relevant features of interest to make a synthetic target feature, and give it to the rule learning algorithms. After a set of rules is derived, some postprocessing steps are taken to make the ruleset more compact and easier to understand by removing some uninteresting or redundant rules. We conducted a set of experiments of mining our traffic accident data in both unsupervised mode and supervised mode for comparison of these rule based learning algorithms. Experiments with the traffic accident data reveals that the association rule learning, in its pure unsupervised mode, can discover some hidden relationship among the features. Under supervised learning setting with combinatorial target feature, however, the subgroup discovery method finds good rules much more easily than the association rule learning method that requires a lot of efforts to tune the parameters.

An Efficient Method for Detecting and Extracting Nomological Networks

The accumulated literature base in the behavioral sciences represents the IS discipline’s greatest source of knowledge, and yet the same literature has grown beyond human comprehension. An experiment is conducted showing the inability of experts to retrieve relevant constructs using full-text search. To address this inability to access the body of theoretical behavioral science research we propose a novel IT artifact built on an information extraction approach to nomological network discovery. Based on the design science paradigm we develop a three-step process for extraction and assembly of nomological networks proceeding through article download, hypothesis extraction, variable extraction, and finally to variable integration. Rule-based vs. machine learning algorithms are evaluated and compared to determine the best approach for the extraction steps. A dataset of all the relevant behavioral studies from two top journals in Information Systems and Psychology is used to evaluate the approach in comparison to expert decisions, leading into a discussion of limitations and possible extensions.

Dynamic traffic steering based on fuzzy Q-Learning approach in a multi-RAT multi-layer wireless network

The infrastructure of current cellular networks must evolve to cope with the increasing demand for mobile-broadband services. Heterogeneous networks are an attractive solution for operators to expand network capacity, based on deploying different Radio Access Technologies, cell sizes and carrier frequencies in the same environment. As a result, operators gain flexibility to distribute traffic across the different networks (or layers) in order to make a more efficient use of resources and enhance network performance. In this work, a dynamic traffic steering technique in multi-RAT multi-layer wireless networks is proposed. In particular, a fuzzy rule-based reinforcement learning algorithm modifies handover parameters according to a specific policy set by the operator, which typically searches for a trade-off between key performance indicators. Results show that the proposed optimization algorithm provides good flexibility to support different policies by simply adjusting some weighting factors. In addition, the Q-Learning algorithm is shown as an effective solution to adapt the network to context variations, such as those produced in the user spatial distribution.

Pronunciation prediction with Default&Refine

The Default&Refine algorithm is a new rule-based learning algorithm that was developed as an accurate and efficient pronunciation prediction mechanism for speech processing systems. The algorithm exhibits a number of attractive properties including rapid generalisation from small training sets, good asymptotic accuracy, robustness to noise in the training data, and the production of compact rule sets. We describe the Default&Refine algorithm in detail and demonstrate its performance on two benchmarked pronunciation databases (the English OALD and Flemish FONILEX pronunciation dictionaries) as well as a newly-developed Afrikaans pronunciation dictionary. We find that the algorithm learns more efficiently (achieves higher accuracy on smaller data sets) than any of the alternative pronunciation prediction algorithms considered. In addition, we demonstrate the ability of the algorithm to generate an arbitrarily small rule set in such a way that the trade-off between rule set size and accuracy is well controlled. A conceptual comparison with alternative algorithms (including Dynamically Expanding Context, Transformation-Based Learning and Pronunciation by Analogy) clarifies the competitive performance obtained with Default&Refine.

Unstructured Object Recognition using Morphological Learning

A technique of object recognition which can detect absence or presence of objects of interest without making explicit use of their underlying geometric structure is deemed suitable for many practical applications. In this work, a method of recognising unstructured objects has been presented, wherein several gray patterns are input as examples to a morphological rule-based learning algorithm. The output of the algorithm are the corresponding gray structuring elements capable of recognising patterns in query images. The learning is carried out offline before recognition of the queries. The technique has been tested to identify fuel pellet surface imperfections. Robustness wrt intensity, orientation, and shape variations of the query patterns is built into the method. Moreover, simplicity of the recognition process leading to reduced computational time makes the method attractive to solve many practical problems.

Combining statistical and reinforcement learning in rule-based classification

BYPASS is a rule-based learning algorithm designed loosely after John Holland’s Classifier System (CS) idea. In essence, CSs are incremental data processors that attempt to (self-)organize a population of rules under the guidance of certain reinforcement policy. BYPASS uses a reinforcement scheme based on predictive scoring and handles conditions similar to those in standard classification trees. However, these conditions are allowed to overlap; this makes it possible to base individual predictions on several rules (just like in committees of trees). Moreover, conditions are supplied with simple bayesian predictive distributions that evolve as data are processed. The paper presents the details of the algorithm and discusses empirical results suggesting that statistical and reinforcement-based learning blend together in interesting and useful ways.

Resolution of pattern recognition problems using a hybrid Genetic/Random Neural Network learning algorithm

Gelenbe has proposed a neural network, called a Random Neural Network, which calculates the probability of activation of the neurons in the network. In this paper, we propose to solve the patterns recognition problem using a hybrid Genetic/Random Neural Network learning algorithm. The hybrid algorithm trains the Random Neural Network by integrating a genetic algorithm with the gradient descent rule-based learning algorithm of the Random Neural Network. This hybrid learning algorithm optimises the Random Neural Network on the basis of its topology and its weights distribution. We apply the hybrid Genetic/Random Neural Network learning algorithm to two pattern recognition problems. The first one recognises or categorises alphabetic characters, and the second recognises geometric figures. We show that this model can efficiently work as associative memory. We can recognise pattern arbitrary images with this algorithm, but the processing time increases rapidly.