Scott Dick

Toward complex fuzzy logic

Complex fuzzy logic is a postulated logic system that is isomorphic to the complex fuzzy sets recently described in a previous paper. This concept is analogous to the many-valued logics that are isomorphic to type-1 fuzzy sets, commonly known as fuzzy logic. As with fuzzy logics, a complex fuzzy logic would be defined by particular choices of the conjunction, disjunction and complement operators. In this paper, an important assertion from a previous paper, that only the modulus of a complex fuzzy membership should be considered in set theoretic (or logical) operations, is examined. A more general mathematical formulation (the property of rotational invariance) is proposed for this assertion, and the impact of this property on the form of complex fuzzy logic operations is examined. All complex fuzzy logics based on the modulus of a vector are shown to be rotationally invariant. The case of complex fuzzy logics that are not rotationally invariant is examined using the framework of vector logic. A candidate conjunction operator was identified, and the existence of a dual disjunction was proven. Finally, a discussion on the possible applications of complex fuzzy logic focuses on the phenomenon of regularity as a possible fuzzification of stationarity.

Consumer trust in e-commerce web sites: A meta-study

Trust is at once an elusive, imprecise concept, and a critical attribute that must be engineered into e-commerce systems. Trust conveys a vast number of meanings, and is deeply dependent upon context. The literature on engineering trust into e-commerce systems reflects these ambiguous meanings; there are a large number of articles, but there is as yet no clear theoretical framework for the investigation of trust in e-commerce. E-commerce, however, is predicated on trust; indeed, any e-commerce vendor that fails to establish a trusting relationship with their customers is doomed. There is a very clear need for specific guidance on e-commerce system attributes and business operations that will effectively promote consumer trust. To address this need, we have conducted a meta-study of the empirical literature on trust in e-commerce systems. This area of research is still immature, and hence our meta-analysis is qualitative rather than quantitative. We identify the major theoretical frameworks that have been proposed in the literature, and propose a qualitative model incorporating the various factors that have been empirically found to influence consumer trust in e-commerce. As this model is too complex to be of practical use, we explore subsets of this model that have the strongest support in the literature, and discuss the implications of this model for Web site design. Finally, we outline key conceptual and methodological needs for future work on this topic.

ANCFIS: A Neurofuzzy Architecture Employing Complex Fuzzy Sets

Complex fuzzy sets (CFSs) are an extension of type-1 fuzzy sets in which the membership of an object to the set is a value from the unit disc of the complex plane. Although there has been considerable progress made in determining the properties of CFSs and complex fuzzy logic, there has yet to be any practical application of this concept. We present the adaptive neurocomplex-fuzzy-inferential system (ANCFIS), which is the first neurofuzzy system architecture to implement complex fuzzy rules (and, in particular, the signature property of rule interference). We have applied this neurofuzzy system to the domain of time-series forecasting, which is an important machine-learning problem. We find that ANCFIS performs well in one synthetic and five real-world forecasting problems and is also very parsimonious. Experimental comparisons show that ANCFIS is comparable with existing approaches on our five datasets. This work demonstrates the utility of complex fuzzy logic on real-world problems.

Applying Novel Resampling Strategies To Software Defect Prediction

Due to the tremendous complexity and sophistication of software, improving software reliability is an enormously difficult task. We study the software defect prediction problem, which focuses on predicting which modules will experience a failure during operation. Numerous studies have applied machine learning to software defect prediction; however, skewness in defect-prediction datasets usually undermines the learning algorithms. The resulting classifiers will often never predict the faulty minority class. This problem is well known in machine learning and is often referred to as learning from unbalanced datasets. We examine stratification, a widely used technique for learning unbalanced data that has received little attention in software defect prediction. Our experiments are focused on the SMOTE technique, which is a method of over-sampling minority-class examples. Our goal is to determine if SMOTE can improve recognition of defect-prone modules, and at what cost. Our experiments demonstrate that after SMOTE resampling, we have a more balanced classification. We found an improvement of at least 23% in the average geometric mean classification accuracy on four benchmark datasets.

Detecting visually similar Web pages: Application to phishing detection

We propose a novel approach for detecting visual similarity between two Web pages. The proposed approach applies Gestalt theory and considers a Web page as a single indivisible entity. The concept of supersignals, as a realization of Gestalt principles, supports our contention that Web pages must be treated as indivisible entities. We objectify, and directly compare, these indivisible supersignals using algorithmic complexity theory. We illustrate our approach by applying it to the problem of detecting phishing scams. Via a large-scale, real-world case study, we demonstrate that 1) our approach effectively detects similar Web pages; and 2) it accuractely distinguishes legitimate and phishing pages.

Data mining in software metrics databases

We investigate the use of data mining for the analysis of software metric databases, and some of the issues in this application domain. Software metrics are collected at various phases of the software development process, in order to monitor and control the quality of a software product. However, software quality control is complicated by the complex relationship between these metrics and the attributes of a software development process. Data mining has been proposed as a potential technology for supporting and enhancing our understanding of software metrics and their relationship to software quality. In this paper, we use fuzzy clustering to investigate three datasets of software metrics, along with the larger issue of whether supervised or unsupervised learning is more appropriate for software engineering problems. While our findings generally confirm the known linear relationship between metrics and change rates, some interesting behaviors are noted. In addition, our results partly contradict earlier studies that only used correlation analysis to investigate these datasets. These results illustrate how intelligent technologies can augment traditional statistical inference in software quality control.

A Survey and Analysis of the P3P Protocol's Agents, Adoption, Maintenance, and Future

In this paper, we survey the adoption of the platform for privacy preferences protocol (P3P) on Internet Web sites to determine if P3P is a growing or stagnant technology. We conducted a pilot survey in February 2005 and our full survey in November 2005. We compare the results from these two surveys and the previous (July 2003) survey of P3P adoption. In general, we find that P3P adoption is stagnant, and errors in P3P documents are a regular occurrence. In addition, very little maintenance of P3P policies is apparent. These observations call into question P3Ps viability as an online privacy-enhancing technology. Our survey exceeds other previous surveys in our use of both detailed statistical analysis and scope; our February pilot survey analyzed more than 23,000 unique Web sites, and our full survey in November 2005 analyzed more than 100,000 unique Web sites.

Highly scalable and robust rule learner: performance evaluation and comparison

Business intelligence and bioinformatics applications increasingly require the mining of datasets consisting of millions of data points, or crafting real-time enterprise-level decision support systems for large corporations and drug companies. In all cases, there needs to be an underlying data mining system, and this mining system must be highly scalable. To this end, we describe a new rule learner called DataSqueezer. The learner belongs to the family of inductive supervised rule extraction algorithms. DataSqueezer is a simple, greedy, rule builder that generates a set of production rules from labeled input data. In spite of its relative simplicity, DataSqueezer is a very effective learner. The rules generated by the algorithm are compact, comprehensible, and have accuracy comparable to rules generated by other state-of-the-art rule extraction algorithms. The main advantages of DataSqueezer are very high efficiency, and missing data resistance. DataSqueezer exhibits log-linear asymptotic complexity with the number of training examples, and it is faster than other state-of-the-art rule learners. The learner is also robust to large quantities of missing data, as verified by extensive experimental comparison with the other learners. DataSqueezer is thus well suited to modern data mining and business intelligence tasks, which commonly involve huge datasets with a large fraction of missing data.

On Pythagorean and Complex Fuzzy Set Operations

Complex fuzzy logic is a new multivalued logic system that has emerged in the last decade. At this time, there are a limited number of known instances of complex fuzzy logic, and only a partial exploration of their properties. There has also been relatively little progress in developing interpretations of complex-valued membership grades. In this paper, we address both problems by examining the recently developed Pythagorean fuzzy sets (a generalization of intuitionistic fuzzy sets). We first characterize two lattices that have been suggested for Pythagorean fuzzy sets and then extend these results to the unit disc of the complex plane. We thereby identify two new complete, distributive lattices over the unit disc, and explore interpretations of them based on fuzzy antonyms and negations.

CRYSTALP2: sequence-based protein crystallization propensity prediction.

BackgroundCurrent protocols yield crystals for <30% of known proteins, indicating that automatically identifying crystallizable proteins may improve high-throughput structural genomics efforts. We introduce CRYSTALP2, a kernel-based method that predicts the propensity of a given protein sequence to produce diffraction-quality crystals. This method utilizes the composition and collocation of amino acids, isoelectric point, and hydrophobicity, as estimated from the primary sequence, to generate predictions. CRYSTALP2 extends its predecessor, CRYSTALP, by enabling predictions for sequences of unrestricted size and provides improved prediction quality.ResultsA significant majority of the collocations used by CRYSTALP2 include residues with high conformational entropy, or low entropy and high potential to mediate crystal contacts; notably, such residues are utilized by surface entropy reduction methods. We show that the collocations provide complementary information to the hydrophobicity and isoelectric point. Tests on four datasets show that CRYSTALP2 outperforms several existing sequence-based predictors (CRYSTALP, OB-score, and SECRET). CRYSTALP2s accuracy, MCC, and AROC range between 69.3 and 77.5%, 0.39 and 0.55, and 0.72 and 0.79, respectively. Our predictions are similar in quality and are complementary to the predictions of the most recent ParCrys and XtalPred methods. Our results also suggest that, as work in protein crystallization continues (thereby enlarging the population of proteins with known crystallization propensities), the prediction quality of the CRYSTALP2 method should increase. The prediction model and the datasets used in this contribution can be downloaded from http://biomine.ece.ualberta.ca/CRYSTALP2/CRYSTALP2.html.ConclusionCRYSTALP2 provides relatively accurate crystallization propensity predictions for a given protein chain that either outperform or complement the existing approaches. The proposed method can be used to support current efforts towards improving the success rate in obtaining diffraction-quality crystals.