Determining process model precision and generalization with weighted artificial negative events

Process mining encompasses the research area which is concerned with knowledge discovery from event logs. One common process mining task focuses on conformance checking, comparing discovered or designed process models with actual real-life behavior as captured in event logs in order to assess the “goodness” of the process model. This paper introduces a novel conformance checking method to measure how well a process model performs in terms of precision and generalization with respect to the actual executions of a process as recorded in an event log. Our approach differs from related work in the sense that we apply the concept of so-called weighted artificial negative events towards conformance checking, leading to more robust results, especially when dealing with less complete event logs that only contain a subset of all possible process execution behavior. In addition, our technique offers a novel way to estimate a process model’s ability to generalize. Existing literature has focused mainly on the fitness (recall) and precision (appropriateness) of process models, whereas generalization has been much more difficult to estimate. The described algorithms are implemented in a number of ProM plugins, and a Petri net conformance checking tool was developed to inspect process model conformance in a visual manner.

Simplified computation and generalization of the refined process structure tree

A business process is often modeled using some kind of a directed flow graph, which we call a workflow graph. The Refined Process Structure Tree (RPST) is a technique for workflow graph parsing, i.e., for discovering the structure of a workflow graph, which has various applications. In this paper, we provide two improvements to the RPST. First, we propose an alternative way to compute the RPST that is simpler than the one developed originally. In particular, the computation reduces to constructing the tree of the triconnected components of a workflow graph in the special case when every node has at most one incoming or at most one outgoing edge. Such graphs occur frequently in applications. Secondly, we extend the applicability of the RPST. Originally, the RPST was applicable only to graphs with a single source and single sink such that the completed version of the graph is biconnected. We lift both restrictions. Therefore, the RPST is then applicable to arbitrary directed graphs such that every node is on a path from some source to some sink. This includes graphs with multiple sources and/or sinks and disconnected graphs.

Mice behaviorally selected for high Pavlovian fear memory exhibit fear generalization, HPA-axis alterations and increased basal activity in cortico-limbic circuits

Biological factors underlying individual variability in fearfulness and anxiety have important implications for stress-related psychiatric illness including PTSD and major depression. Using an advanced intercross line (AIL) derived from C57BL/6 and DBA/2J mouse strains and behavioral selection over 3 generations, we established two lines exhibiting High or Low fear behavior after fear conditioning. Across the selection generations, the two lines showed clear differences in training and tests for contextual and conditioned fear. Before fear conditioning training, there were no differences between lines in baseline freezing to a novel context. However, after fear conditioning High line mice demonstrated pronounced freezing in a new context suggestive of poor context discrimination. Fear generalization was not restricted to contextual fear. High fear mice froze to a novel acoustic stimulus while freezing in the Low line did not increase over baseline. Enhanced fear learning and generalization are consistent with transgenic and pharmacological disruption of the hypothalamic-pituitary-adrenal axis (HPA-axis) (Brinks, 2009, Thompson, 2004, Kaouane, 2012). To determine whether there were differences in HPA-axis regulation between the lines, morning urine samples were collected to measure basal corticosterone. Levels of secreted corticosterone in the circadian trough were analyzed by corticosterone ELISA. High fear mice were found to have higher basal corticosterone levels than low line animals. Examination of hormonal stress response components by qPCR revealed increased expression of CRH mRNA and decreased mRNA for MR and CRHR1 in hypothalamus of high fear mice. These alterations may contribute to both the behavioral phenotype and higher basal corticosterone in High fear mice. To determine basal brain activity in vivo in High and Low fear mice we used manganese-enhanced magnetic resonance imaging (MEMRI). Analysis revealed a pattern of basal brain activity made up of amygdala, cortical and hippocampal circuits that was elevated in the High line. Ongoing studies also seek to determine the relative balance of excitatory and inhibitory tone in the amygdala and hippocampus and the neuronal structure of its neurons. While these heterogeneous lines are selected on fear memory expression, HPA-axis alterations and differences in hippocampal activity segregate with the behavioral phenotypes. These differences are detectable in a basal state strongly suggesting these are biological traits underlying the behavioral phenotype (Johnson et al, 2011).

A study of Chinese private passenger cars based on BP neural network

This study employs BP neural network to simulate the development of Chinese private passenger cars. Considering the uncertain and complex environment for the development of private passenger cars, indicators of economy, population, price, infrastructure, income, energy and some other fields which have major impacts on it are selected at first. The network is proved to be operable to simulate the progress of chinese private passenger cars after modeling, training and generalization test. Based on the BP neural network model, sensitivity analysis of each indicator is carried on and shows that the sensitivity coefficients of fuel price change suddenly. This special phenomenon reveals that the development of Chinese private passenger cars may be seriously affected by the recent high fuel price. This finding is also consistent with facts and figures

Revisiting the conceptualization and measurement of service quality

Since the 1980s, industries and researchers have sought to better understand the quality of services due to the rise in their importance (Brogowicz, Delene and Lyth 1990). More recent developments with online services, coupled with growing recognition of service quality (SQ) as a key contributor to national economies and as an increasingly important competitive differentiator, amplify the need to revisit our understanding of SQ and its measurement. Although ‘SQ’ can be broadly defined as “a global overarching judgment or attitude relating to the overall excellence or superiority of a service” (Parasuraman, Berry and Zeithaml 1988), the term has many interpretations. There has been considerable progress on how to measure SQ perceptions, but little consensus has been achieved on what should be measured. There is agreement that SQ is multi-dimensional, but little agreement as to the nature or content of these dimensions (Brady and Cronin 2001). For example, within the banking sector, there exist multiple SQ models, each consisting of varying dimensions. The existence of multiple conceptions and the lack of a unifying theory bring the credibility of existing conceptions into question, and beg the question of whether it is possible at some higher level to define SQ broadly such that it spans all service types and industries. This research aims to explore the viability of a universal conception of SQ, primarily through a careful re-visitation of the services and SQ literature. The study analyses the strengths and weaknesses of the highly regarded and widely used global SQ model (SERVQUAL) which reflects a single-level approach to SQ measurement. The SERVQUAL model states that customers evaluate SQ (of each service encounter) based on five dimensions namely reliability, assurance, tangibles, empathy and responsibility. SERVQUAL, however, failed to address what needs to be reliable, assured, tangible, empathetic and responsible. This research also addresses a more recent global SQ model from Brady and Cronin (2001); the B&C (2001) model, that has potential to be the successor of SERVQUAL in that it encompasses other global SQ models and addresses the ‘what’ questions that SERVQUAL didn’t. The B&C (2001) model conceives SQ as being multidimensional and multi-level; this hierarchical approach to SQ measurement better reflecting human perceptions. In-line with the initial intention of SERVQUAL, which was developed to be generalizable across industries and service types, this research aims to develop a conceptual understanding of SQ, via literature and reflection, that encompasses the content/nature of factors related to SQ; and addresses the benefits and weaknesses of various SQ measurement approaches (i.e. disconfirmation versus perceptions-only). Such understanding of SQ seeks to transcend industries and service types with the intention of extending our knowledge of SQ and assisting practitioners in understanding and evaluating SQ. The candidate’s research has been conducted within, and seeks to contribute to, the ‘IS-Impact’ research track of the IT Professional Services (ITPS) Research Program at QUT. The vision of the track is “to develop the most widely employed model for benchmarking Information Systems in organizations for the joint benefit of research and practice.” The ‘IS-Impact’ research track has developed an Information Systems (IS) success measurement model, the IS-Impact Model (Gable, Sedera and Chan 2008), which seeks to fulfill the track’s vision. Results of this study will help future researchers in the ‘IS-Impact’ research track address questions such as: • Is SQ an antecedent or consequence of the IS-Impact model or both? • Has SQ already been addressed by existing measures of the IS-Impact model? • Is SQ a separate, new dimension of the IS-Impact model? • Is SQ an alternative conception of the IS? Results from the candidate’s research suggest that SQ dimensions can be classified at a higher level which is encompassed by the B&C (2001) model’s 3 primary dimensions (interaction, physical environment and outcome). The candidate also notes that it might be viable to re-word the ‘physical environment quality’ primary dimension to ‘environment quality’ so as to better encompass both physical and virtual scenarios (E.g: web sites). The candidate does not rule out the global feasibility of the B&C (2001) model’s nine sub-dimensions, however, acknowledges that more work has to be done to better define the sub-dimensions. The candidate observes that the ‘expertise’, ‘design’ and ‘valence’ sub-dimensions are supportive representations of the ‘interaction’, physical environment’ and ‘outcome’ primary dimensions respectively. The latter statement suggests that customers evaluate each primary dimension (or each higher level of SQ classification) namely ‘interaction’, physical environment’ and ‘outcome’ based on the ‘expertise’, ‘design’ and ‘valence’ sub-dimensions respectively. The ability to classify SQ dimensions at a higher level coupled with support for the measures that make up this higher level, leads the candidate to propose the B&C (2001) model as a unifying theory that acts as a starting point to measuring SQ and the SQ of IS. The candidate also notes, in parallel with the continuing validation and generalization of the IS-Impact model, that there is value in alternatively conceptualizing the IS as a ‘service’ and ultimately triangulating measures of IS SQ with the IS-Impact model. These further efforts are beyond the scope of the candidate’s study. Results from the candidate’s research also suggest that both the disconfirmation and perceptions-only approaches have their merits and the choice of approach would depend on the objective(s) of the study. Should the objective(s) be an overall evaluation of SQ, the perceptions-only approached is more appropriate as this approach is more straightforward and reduces administrative overheads in the process. However, should the objective(s) be to identify SQ gaps (shortfalls), the (measured) disconfirmation approach is more appropriate as this approach has the ability to identify areas that need improvement.

Novel educational training program for community pharmacists

There is little evidence that workshops alone have a lasting impact on the day-to-day practice of participants. The current paper examined a strategy to increase generalization and maintenance of skills in the natural environment using pseudo-patients and immediate performance feedback to reinforce skills acquisition. A random half of pharmacies (N=30) took part in workshop training aimed at optimizing consumers' use of nonprescription analgesic products. Pharmacies in the training group also received performance feedback on their adherence to the recommended protocol. Feedback occurred immediately after a pseudo-patient visit in which confederates posed as purchasers of analgesics, and combined positive and corrective elements. Trained pharmacists were significantly more accurate at identifying people who misused the medication (P<0.001). The trained pharmacists were more likely than controls to use open-ended questions (P<0.001), assess readiness to change problematic use (P <0.001), and to deliver a brief intervention that was tailored to the person's commitment to alter his/her usage (P <0.001). Participants responded to the feedback positively. Results were consistent with the hypothesis that when workshop is combined with on-site performance feedback, it enhances practitioners' adherence to protocols in the natural setting.

Decision support system for contractor pre-qualification : artificial neural network model

The selection criteria for contractor pre-qualification are characterized by the co-existence of both quantitative and qualitative data. The qualitative data is non-linear, uncertain and imprecise. An ideal decision support system for contractor pre-qualification should have the ability of handling both quantitative and qualitative data, and of mapping the complicated nonlinear relationship of the selection criteria, such that rational and consistent decisions can be made. In this research paper, an artificial neural network model was developed to assist public clients identifying suitable contractors for tendering. The pre-qualification criteria (variables) were identified for the model. One hundred and twelve real pre-qualification cases were collected from civil engineering projects in Hong Kong, and eighty-eight hypothetical pre-qualification cases were also generated according to the “If-then” rules used by professionals in the pre-qualification process. The results of the analysis totally comply with current practice (public developers in Hong Kong). Each pre-qualification case consisted of input ratings for candidate contractors’ attributes and their corresponding pre-qualification decisions. The training of the neural network model was accomplished by using the developed program, in which a conjugate gradient descent algorithm was incorporated for improving the learning performance of the network. Cross-validation was applied to estimate the generalization errors based on the “re-sampling” of training pairs. The case studies show that the artificial neural network model is suitable for mapping the complicated nonlinear relationship between contractors’ attributes and their corresponding pre-qualification (disqualification) decisions. The artificial neural network model can be concluded as an ideal alternative for performing the contractor pre-qualification task.

A Cultural Science Manifesto (in Mandarin)

We propose that a general analytic framework for cultural science can be constructed as a generalization of the generic micro meso macro framework proposed by Dopfer and Potts (2008). This paper outlines this argument along with some implications for the creative industries research agenda.

Econometric analysis of stuctural systems with permanent and transitory shocks

This paper considers the implications of the permanent/transitory decomposition of shocks for identification of structural models in the general case where the model might contain more than one permanent structural shock. It provides a simple and intuitive generalization of the influential work of Blanchard and Quah [1989. The dynamic effects of aggregate demand and supply disturbances. The American Economic Review 79, 655–673], and shows that structural equations with known permanent shocks cannot contain error correction terms, thereby freeing up the latter to be used as instruments in estimating their parameters. The approach is illustrated by a re-examination of the identification schemes used by Wickens and Motto [2001. Estimating shocks and impulse response functions. Journal of Applied Econometrics 16, 371–387], Shapiro and Watson [1988. Sources of business cycle fluctuations. NBER Macroeconomics Annual 3, 111–148], King et al. [1991. Stochastic trends and economic fluctuations. American Economic Review 81, 819–840], Gali [1992. How well does the ISLM model fit postwar US data? Quarterly Journal of Economics 107, 709–735; 1999. Technology, employment, and the business cycle: Do technology shocks explain aggregate fluctuations? American Economic Review 89, 249–271] and Fisher [2006. The dynamic effects of neutral and investment-specific technology shocks. Journal of Political Economy 114, 413–451].

Fingerprint analysis using wavelet transform with application to compression and feature extraction

The main goal of this research is to design an efficient compression al~ gorithm for fingerprint images. The wavelet transform technique is the principal tool used to reduce interpixel redundancies and to obtain a parsimonious representation for these images. A specific fixed decomposition structure is designed to be used by the wavelet packet in order to save on the computation, transmission, and storage costs. This decomposition structure is based on analysis of information packing performance of several decompositions, two-dimensional power spectral density, effect of each frequency band on the reconstructed image, and the human visual sensitivities. This fixed structure is found to provide the "most" suitable representation for fingerprints, according to the chosen criteria. Different compression techniques are used for different subbands, based on their observed statistics. The decision is based on the effect of each subband on the reconstructed image according to the mean square criteria as well as the sensitivities in human vision. To design an efficient quantization algorithm, a precise model for distribution of the wavelet coefficients is developed. The model is based on the generalized Gaussian distribution. A least squares algorithm on a nonlinear function of the distribution model shape parameter is formulated to estimate the model parameters. A noise shaping bit allocation procedure is then used to assign the bit rate among subbands. To obtain high compression ratios, vector quantization is used. In this work, the lattice vector quantization (LVQ) is chosen because of its superior performance over other types of vector quantizers. The structure of a lattice quantizer is determined by its parameters known as truncation level and scaling factor. In lattice-based compression algorithms reported in the literature the lattice structure is commonly predetermined leading to a nonoptimized quantization approach. In this research, a new technique for determining the lattice parameters is proposed. In the lattice structure design, no assumption about the lattice parameters is made and no training and multi-quantizing is required. The design is based on minimizing the quantization distortion by adapting to the statistical characteristics of the source in each subimage. 11 Abstract Abstract Since LVQ is a multidimensional generalization of uniform quantizers, it produces minimum distortion for inputs with uniform distributions. In order to take advantage of the properties of LVQ and its fast implementation, while considering the i.i.d. nonuniform distribution of wavelet coefficients, the piecewise-uniform pyramid LVQ algorithm is proposed. The proposed algorithm quantizes almost all of source vectors without the need to project these on the lattice outermost shell, while it properly maintains a small codebook size. It also resolves the wedge region problem commonly encountered with sharply distributed random sources. These represent some of the drawbacks of the algorithm proposed by Barlaud [26). The proposed algorithm handles all types of lattices, not only the cubic lattices, as opposed to the algorithms developed by Fischer [29) and Jeong [42). Furthermore, no training and multiquantizing (to determine lattice parameters) is required, as opposed to Powell's algorithm [78). For coefficients with high-frequency content, the positive-negative mean algorithm is proposed to improve the resolution of reconstructed images. For coefficients with low-frequency content, a lossless predictive compression scheme is used to preserve the quality of reconstructed images. A method to reduce bit requirements of necessary side information is also introduced. Lossless entropy coding techniques are subsequently used to remove coding redundancy. The algorithms result in high quality reconstructed images with better compression ratios than other available algorithms. To evaluate the proposed algorithms their objective and subjective performance comparisons with other available techniques are presented. The quality of the reconstructed images is important for a reliable identification. Enhancement and feature extraction on the reconstructed images are also investigated in this research. A structural-based feature extraction algorithm is proposed in which the unique properties of fingerprint textures are used to enhance the images and improve the fidelity of their characteristic features. The ridges are extracted from enhanced grey-level foreground areas based on the local ridge dominant directions. The proposed ridge extraction algorithm, properly preserves the natural shape of grey-level ridges as well as precise locations of the features, as opposed to the ridge extraction algorithm in [81). Furthermore, it is fast and operates only on foreground regions, as opposed to the adaptive floating average thresholding process in [68). Spurious features are subsequently eliminated using the proposed post-processing scheme.

Examination of gaze behaviors under in situ and video simulation task constraints reveals differences in information pickup for perception and action

Gaze and movement behaviors of association football goalkeepers were compared under two video simulation conditions (i.e., verbal and joystick movement responses) and three in situ conditions (i.e., verbal, simplified body movement, and interceptive response). The results showed that the goalkeepers spent more time fixating on information from the penalty kick taker’s movements than ball location for all perceptual judgment conditions involving limited movement (i.e., verbal responses, joystick movement, and simplified body movement). In contrast, an equivalent amount of time was spent fixating on the penalty taker’s relative motions and the ball location for the in situ interception condition, which required the goalkeepers to attempt to make penalty saves. The data suggest that gaze and movement behaviors function differently, depending on the experimental task constraints selected for empirical investigations. These findings highlight the need for research on perceptual— motor behaviors to be conducted in representative experimental conditions to allow appropriate generalization of conclusions to performance environments.

The sample complexity of pattern classification with neural networks: the size of the weights is more important than the size of the network

Sample complexity results from computational learning theory, when applied to neural network learning for pattern classification problems, suggest that for good generalization performance the number of training examples should grow at least linearly with the number of adjustable parameters in the network. Results in this paper show that if a large neural network is used for a pattern classification problem and the learning algorithm finds a network with small weights that has small squared error on the training patterns, then the generalization performance depends on the size of the weights rather than the number of weights. For example, consider a two-layer feedforward network of sigmoid units, in which the sum of the magnitudes of the weights associated with each unit is bounded by A and the input dimension is n. We show that the misclassification probability is no more than a certain error estimate (that is related to squared error on the training set) plus A3 √((log n)/m) (ignoring log A and log m factors), where m is the number of training patterns. This may explain the generalization performance of neural networks, particularly when the number of training examples is considerably smaller than the number of weights. It also supports heuristics (such as weight decay and early stopping) that attempt to keep the weights small during training. The proof techniques appear to be useful for the analysis of other pattern classifiers: when the input domain is a totally bounded metric space, we use the same approach to give upper bounds on misclassification probability for classifiers with decision boundaries that are far from the training examples.

A unifying view of multiple kernel learning

Recent research on multiple kernel learning has lead to a number of approaches for combining kernels in regularized risk minimization. The proposed approaches include different formulations of objectives and varying regularization strategies. In this paper we present a unifying optimization criterion for multiple kernel learning and show how existing formulations are subsumed as special cases. We also derive the criterion’s dual representation, which is suitable for general smooth optimization algorithms. Finally, we evaluate multiple kernel learning in this framework analytically using a Rademacher complexity bound on the generalization error and empirically in a set of experiments.

Shifting : one-inclusion mistake bounds and sample compression

We present new expected risk bounds for binary and multiclass prediction, and resolve several recent conjectures on sample compressibility due to Kuzmin and Warmuth. By exploiting the combinatorial structure of concept class F, Haussler et al. achieved a VC(F)/n bound for the natural one-inclusion prediction strategy. The key step in their proof is a d = VC(F) bound on the graph density of a subgraph of the hypercube—oneinclusion graph. The first main result of this paper is a density bound of n [n−1 <=d-1]/[n <=d] < d, which positively resolves a conjecture of Kuzmin and Warmuth relating to their unlabeled Peeling compression scheme and also leads to an improved one-inclusion mistake bound. The proof uses a new form of VC-invariant shifting and a group-theoretic symmetrization. Our second main result is an algebraic topological property of maximum classes of VC-dimension d as being d contractible simplicial complexes, extending the well-known characterization that d = 1 maximum classes are trees. We negatively resolve a minimum degree conjecture of Kuzmin and Warmuth—the second part to a conjectured proof of correctness for Peeling—that every class has one-inclusion minimum degree at most its VCdimension. Our final main result is a k-class analogue of the d/n mistake bound, replacing the VC-dimension by the Pollard pseudo-dimension and the one-inclusion strategy by its natural hypergraph generalization. This result improves on known PAC-based expected risk bounds by a factor of O(logn) and is shown to be optimal up to an O(logk) factor. The combinatorial technique of shifting takes a central role in understanding the one-inclusion (hyper)graph and is a running theme throughout.

Shifting : one-inclusion mistake bounds and sample compression

We present new expected risk bounds for binary and multiclass prediction, and resolve several recent conjectures on sample compressibility due to Kuzmin and Warmuth. By exploiting the combinatorial structure of concept class F, Haussler et al. achieved a VC(F)/n bound for the natural one-inclusion prediction strategy. The key step in their proof is a d=VC(F) bound on the graph density of a subgraph of the hypercube—one-inclusion graph. The first main result of this report is a density bound of n∙choose(n-1,≤d-1)/choose(n,≤d) < d, which positively resolves a conjecture of Kuzmin and Warmuth relating to their unlabeled Peeling compression scheme and also leads to an improved one-inclusion mistake bound. The proof uses a new form of VC-invariant shifting and a group-theoretic symmetrization. Our second main result is an algebraic topological property of maximum classes of VC-dimension d as being d-contractible simplicial complexes, extending the well-known characterization that d=1 maximum classes are trees. We negatively resolve a minimum degree conjecture of Kuzmin and Warmuth—the second part to a conjectured proof of correctness for Peeling—that every class has one-inclusion minimum degree at most its VC-dimension. Our final main result is a k-class analogue of the d/n mistake bound, replacing the VC-dimension by the Pollard pseudo-dimension and the one-inclusion strategy by its natural hypergraph generalization. This result improves on known PAC-based expected risk bounds by a factor of O(log n) and is shown to be optimal up to a O(log k) factor. The combinatorial technique of shifting takes a central role in understanding the one-inclusion (hyper)graph and is a running theme throughout