Use of conditional rule structure to automate clinical decision support: a comparison of artificial intelligence and deterministic programming techniques

A rule-based computer system was developed to perform clinical decision-making support within a medical information system, oncology practice, and clinical research. This rule-based system, which has been programmed using deterministic rules, possesses features of generalizability, modularity of structure, convenience in rule acquisition, explanability, and utility for patient care and teaching, features which have been identified as advantages of artificial intelligence (AI) rule-based systems. Formal rules are primarily represented as conditional statements; common conditions and actions are stored in system dictionaries so that they can be recalled at any time to form new decision rules. Important similarities and differences exist in the structure of this system and clinical computer systems utilizing artificial intelligence (AI) production rule techniques. The non-AI rule-based system possesses advantages in cost and ease of implementation. The degree to which significant medical decision problems can be solved by this technique remains uncertain as does whether the more complex AI methodologies will be required.     

上下文感知系统中的规则生成与匹配算法

针对现有上下文感知系统中的规则主要依靠开发者或用户手工定义的问题,提出了一种基于粗糙集理论的自动规则生成方法.该方法将上下文感知系统视为一种决策信息系统,并利用可辨识矩阵对上下文信息加以约简,进而自动生成规则.由于可供使用的数据有限,所生成的规则无法完全覆盖上下文的取值范围,因此可能出现找不到与上下文状态相匹配规则的问题.为了解决这一问题,提出了一种基于语义距离的规则匹配算法.最后验证了所提出方法的有效性和效率.     

A genetic algorithm-based rule extraction system

Individual classifiers predict unknown objects. Although, these are usually domain specific, and lack the property of scaling up prediction while handling data sets with huge size and high-dimensionality or imbalance class distribution. This article introduces an accuracy-based learning system called DTGA (decision tree and genetic algorithm) that aims to improve prediction accuracy over any classification problem irrespective to domain, size, dimensionality and class distribution. More specifically, the proposed system consists of two rule inducing phases. In the first phase, a base classifier, C4.5 (a decision tree based rule inducer) is used to produce rules from training data set, whereas GA (genetic algorithm) in the next phase refines them with the aim to provide more accurate and high-performance rules for prediction. The system has been compared with competent non-GA based systems: neural network, Naïve Bayes, rule-based classifier using rough set theory and C4.5 (i.e., the base classifier of DTGA), on a number of benchmark datasets collected from UCI (University of California at Irvine) machine learning repository. Empirical results demonstrate that the proposed hybrid approach provides marked improvement in a number of cases.     

决策树剪枝加强的关联规则分类方法

传统关联规则挖掘在面临分类决策问题时,易出现非频繁规则遗漏、预测精度不高的问题。为得到正确合理且更为完整的规则,提出了一种改进方法 DT-AR(decision tree-association rule algorithm),利用决策树剪枝策略对关联规则集进行补充。该方法利用FP-Growth(frequent pattern growth)算法得到关联规则集,利用C4.5算法构建后剪枝决策树并提取分类规则,在进行置信度迭代筛选后与关联规则集取并集修正,利用置信度作为权重系数采取投票法进行分类。实验结果表明,与传统关联规则挖掘和决策树剪枝方法相比,该方法得到的规则在数据集分类结果上更准确。     

Structured firewall design

A firewall is a security guard placed at the point of entry between a private network and the outside Internet such that all incoming and outgoing packets have to pass through it. The function of a firewall is to examine every incoming or outgoing packet and decide whether to accept or discard it. This function is conventionally specified by a sequence of rules, where rules often conflict. To resolve conflicts, the decision for each packet is the decision of the first rule that the packet matches. The current practice of designing a firewall directly as a sequence of rules suffers from three types of major problems: (1) the consistency problem, which means that it is difficult to order the rules correctly; (2) the completeness problem, which means that it is difficult to ensure thorough consideration for all types of traffic; (3) the compactness problem, which means that it is difficult to keep the number of rules small (because some rules may be redundant and some rules may be combined into one rule).To achieve consistency, completeness, and compactness, we propose a new method called structured firewall design, which consists of two steps. First, one designs a firewall using a firewall decision diagram instead of a sequence of often conflicting rules. Second, a program converts the firewall decision diagram into a compact, yet functionally equivalent, sequence of rules. This method addresses the consistency problem because a firewall decision diagram is conflict-free. It addresses the completeness problem because the syntactic requirements of a firewall decision diagram force the designer to consider all types of traffic. It also addresses the compactness problem because in the second step we use two algorithms (namely FDD reduction and FDD marking) to combine rules together, and one algorithm (namely firewall compaction) to remove redundant rules. Moreover, the techniques and algorithms presented in this paper are extensible to other rule-based systems such as IPsec rules.     

Sequential decision aggregation with social pressure

This paper proposes and characterizes a sequential decision aggregation system consisting of agents performing binary sequential hypothesis testing and a fusion center which collects the individual decisions and reaches the global decision according to some threshold rule. Individual decision makers’ behaviors in the system are influenced by other decision makers, through a model for social pressure; our notion of social pressure is proportional to the ratio of individual decision makers who have already made the decisions. For our proposed model, we obtain the following results: First, we derive a recursive expression for the probabilities of making the correct and wrong global decisions as a function of time, system size, and the global decision threshold. The expression is based on the individual decision makers’ decision probabilities and does not rely on the specific individual decision-making policy. Second, we discuss two specific threshold rules: the fastest rule and the majority rule. By means of a mean-field analysis, we relate the asymptotic performance of the fusion center, as the system size tends to infinity, to the individual decision makers’ decision probability sequence. In addition to theoretical analysis, simulation work is conducted to discuss the speed/accuracy tradeoffs for different threshold rules.     

Orienting rewrite rules with the Knuth–Bendix order

We consider two decision problems related to the Knuth–Bendix order (KBO). The first problem is orientability: given a system of rewrite rules R, does there exist an instance of KBO which orients every ground instance of every rewrite rule in R. The second problem is whether a given instance of KBO orients every ground instance of a given rewrite rule. This problem can also be reformulated as the problem of solving a single ordering constraint for the KBO. We prove that both problems can be solved in the time polynomial in the size of the input. The polynomial-time algorithm for orientability builds upon an algorithm for solving systems of homogeneous linear inequalities over integers. We show that the orientability problem is P-complete. The polynomial-time algorithm for solving a single ordering constraint does not need to solve systems of linear inequalities and can be run in time O(n²). Also we show that if a system is orientable using a real-valued instance of KBO, then it is also orientable using an integer-valued instance of KBO. Therefore, all our results hold both for the integer-valued and the real-valued KBO.     

Small universal simple spiking neural P systems with weights

Spiking neural P systems with weights（WSN P systems,for short） are a new variant of spiking neural P systems,where the rules of a neuron are enabled when the potential of that neuron equals a given value.It is known that WSN P systems are universal by simulating register machines. However,in these universal systems,no bound is considered on the number of neurons and rules. In this work,a restricted variant of WSN P systems is considered,called simple WSN P systems,where each neuron has only one rule. The complexity parameter,the number of neurons,to construct a universal simple WSN P system is investigated. It is proved that there is a universal simple WSN P system with 48 neurons for computing functions; as generator of sets of numbers,there is an almost simple（that is,each neuron has only one rule except that one neuron has two rules） and universal WSN P system with 45 neurons.     

基于GDT的不完整信息系统规则发现

提出了一个基于GDT的从不完整信息系统进行规则发现的方法。该方法利用GDT的思想,通过概括强度、规则置信度和规则强度等概念,充分考虑到数据不完整性和噪音引起的不确定性,在不改变原信息系统大小的前提下,直接从不完整信息系统中得到简洁实用的规则。最后,通过一个例子阐述了该方法的实施过程,并将该方法与提及的其它几种不从不完整信息系统中发现规则的方法进行了分析比较。分析表明该方法是一种新的有效的从不完整信息系统抽取规则的途径。     

Dominance-based Rough Set Classifier without Induction of Decision Rules

Rough Sets Theory is often applied to the task of classification and prediction, in which objects are assigned to some pre-defined decision classes. When the classes are preference-ordered, the process of classification is referred to as sorting. To deal with the specificity of sorting problems an extension of the Classic Rough Sets Approach, called the Dominance-based Rough Sets Approach, was introduced. The final result of the analysis is a set of decision rules induced from what is called rough approximations of decision classes. The main role of the induced decision rules is to discover regularities in the analyzed data set, but the same rules, when combined with a particular classification method, may also be used to classify/sort new objects (i.e. to assign the objects to appropriate classes). There exist many different rule induction strategies, including induction of an exhaustive set of rules. This strategy produces the most comprehensive knowledge base on the analyzed data set, but it requires a considerable amount of computing time, as the complexity of the process is exponential. In this paper we present a shortcut that allows classifying new objects without generating the rules. The presented approach bears some resemblance to the idea of lazy learning.     

基于序列关联规则挖掘的Web日志预测精度研究

目前许多研究关注如何利用序列关联规则预测用户最近的HTTP请求，这些研究主要利用次序信息或时间信息来进行剪枝，以提高预测的精度。该文对不同序列关联规则进行了分析和比较，给出了不同次序信息和时间信息的条件下各种序列模式挖掘算法。并使用实验比较这些算法的预测精度。通过对实验结果的分析，为进一步提高预测的精度指明了方向。     

Unexpected properties and optimum-distributed sensor detectors fordependent observation cases

Optimum-distributed signal detection system design is studied for cases with statistically dependent observations from sensor to sensor. The common parallel architecture is assumed. Here, each sensor sends a decision to a fusion center that determines a final binary decision using a nonrandomized fusion rule. General L sensor cases are considered. A discretized iterative algorithm is suggested that can provide approximate solutions to the necessary conditions for optimum distributed sensor decision rules under a fixed fusion rule. The algorithm is shown to converge in a finite number of iterations, and the solutions obtained are shown to approach the solutions to the original problem, without discretization, as the variable step size shrinks to zero. In the formulation, both binary and multiple-bit sensor decisions cases are considered. Illustrative numerical examples are presented for two-, three-, and four-sensor cases, in which a common random Gaussian signal is to be detected in Gaussian noise     

A logic-based approach to rule induction in expert systems

Rule induction is an important part of learning in expert systems. Rules can help managers make more effective decisions and gain insight into the relationships between decision variables. We present a logic-based approach to rule induction in expert systems which is simple, robust and consistent. We also derive bounds on levels of certainty for combining rules. We apply our approach to the development of rules for the entry decisions of new products. We then discuss how the logic-based approach of rule induction can be used to create a decision support system and the methodology to create such a system.     

基于知识的神经网络在出行方式选择中的应用研究*

针对神经网络和决策树方法在算法上的本质联系和互补优势,将C4.5决策树提取规则的基于知识的神经网络（knowledgebased neural network,KBNN）用于出行方式预测。对居民通勤出行方式选择数据的分析表明,KBNN相比于决策树方法、普通前馈神经网络和多项Logit模型（MNL）有更高的预测精度,方法不仅提高了网络的可解释性,且易于构造、收敛速度更快,实用性较强,为出行方式选择预测提供了新的思路。     

A comparative study of combining multiple enrolled samples for fingerprint verification

In fingerprint verification systems, there are usually multiple (from two to four) enrolled impressions for a same finger. The performance of the systems can be improved by combining these impressions through feature fusion or decision fusion strategy. In this paper, different schemes to combine multiple enrolled impressions are comparatively studied. Experimental results show that a larger improvement can be obtained by using decision fusion scheme than feature fusion. In all decision fusion rules, sum rule outperforms voting rule a little whether using similarity or Neyman-Pearson rule. Based on the observation that the performance of these two strategies can complement each other, we also propose a novel fusion scheme to further combine feature fusion and decision fusion, which can produce an even better result.     

MOSAR: A Multi-Objective Strategy for Hiding Sensitive Association Rules Using Genetic Algorithm

It seems very realistic to find different aspects in a problem solution like rule hiding. Based on this point of view, Availability, Sensitivity, and Conflict are defined as the novel measurements to detect specific transactions in each transactional database in an effective way. At this point, it will be helpful for decision makers to consider such aspects on their solutions to hide sensitive association rules (ARs). Accordingly, the authors put forward a fitness function of genetic algorithm adjusted to the proposed measurements for hiding sensitive rules of the original database. Experimental study shows that the MOSAR algorithm outperforms traditional approach (Decrease Support of Right Hand Side algorithm) in view of reducing ARs in conflict with sensitive ARs, as a side effect in sensitive rule hiding process. Furthermore, this approach is applicable to do its best in banking systems, where database is to be shared through the networks, to protect strategic information after a serious attack.     

A real-time inventory decision system using Western Electric run rules and ARMA control chart

Inventory management is an important area of production control. In 1999, Pfohl et al. [Pfohl, H.-C., Cullmann, O., & Stölzle, W. (1999). Inventory management with statistical process control: Simulation and evaluation. Journal of Business Logistics, 20, 101–120] developed a real-time inventory decision support system by using the individual control charts for monitoring the inventory level (i.e., stock quantity) and the market demand, in which a series of decision rules are provided to help the inventory manager to determine the time and the quantity to order. In the present paper, a real-time inventory decision system is proposed by incorporating Western Electric run rules into the decision rules of the system. Since the data of demand sometimes present a pattern of time series (i.e., autocorrelation may exist in the data of demand), in the proposed decision system the ARMA control chart is employed to monitor the market demand and the individual control chart is used to monitor the inventory level. A simulation study is conducted to investigate the effects of demand pattern and autocorrelation on the proposed inventory decision system and to verify the effectiveness of the system. The index “service level” is selected as the key indicator for the system performance. Based on the results of the simulation study, it is shown that the performance of the proposed inventory decision system is quite consistent with service level always greater than 90% for various demand patterns.     

Hierarchical object-driven action rules

In this work, we present the hierarchical object-driven action rules; a hybrid action rule extraction approach that combines key elements from both the classical action rule mining approach, first proposed by Ra? and Wieczorkowska (2000), and the more recent object-driven action rule extraction approach proposed by Hajja et al. (2012, 2013), to extract action rules from object-driven information systems. Action rules, as defined in Ra? and Wieczorkowska (2000), are actionable tasks that describe possible transitions of instances from one state to another with respect to a distinguished attribute, called the decision attribute. Recently, a new specialized case of action rules, namely object-driven action rules, has been introduced by Hajja et al. (2012, 2013). Object-driven action rules are action rules that are extracted from information systems with temporal and object-based nature. By object-driven information systems, we mean systems that contain multiple observations for each object, in which objects are determined by an attribute that assumingly defines some unique distribution; and by temporally-based information systems, we refer to systems in which each instance is attached to a timestamp that, by definition, must have an intrinsic meaning for each corresponding instance. Though the notion of object-driven and temporal-based action rules had its own successes, some argue that the essence of object-driven assumptions, which is in big part the reason for its effectiveness, are imposing few limitations as well. Object-driven approaches treat entire systems as multi-subsystems for which action rules are extracted from; as a result, more accurate and specific action rules are extracted. However, by doing so, our diverseness of the extracted action rules are much less apparent, compared to the outcome when applying the classical action rule extraction approach, which treats information systems as a whole. For that reason, we propose a hybrid approach which builds a hierarchy of clusters of subsystems; a novel way of clustering through treatments responses similarities is introduced.