The definition of a distance measure plays key role in the evaluation different clustering solutions gene expression profiles. In this empirical study we compare when using Mutual Information (MI) versus use well known Euclidean and Pearson correlation coefficient.Relying on several public datasets, evaluate homogeneity separation scores solutions. It was found that MI yields more significant differentiation among erroneous proposed also used to analyze performance algorithms. A comparative...

Motivation: We propose a new class of variable-order Bayesian network (VOBN) models for the identification transcription factor binding sites (TFBSs). The proposed generalize widely used position weight matrix (PWM) models, Markov and models. In contrast to these where each fixed subset remaining positions is model dependencies, in VOBN subsets may vary based on specific nucleotides observed, which are called context. This flexibility turns out be advantage classification analysis TFBSs, as...

The construction of efficient decision and classification trees is a fundamental task in Big Data analytics which known to be NP-hard. Accordingly, many greedy heuristics were suggested for the decision-trees, but found result local-optimum solutions. In this work we present dual information distance (DID) method that computationally attractive, yet relatively robust noise. DID heuristic selects features by considering both their immediate contribution classification, as well future...

Abstract The increasing use of computerized tools for virtual manufacturing in workstatin design has two main advantages over traditional methods first it enables the designer to examine a large number solutions; and second, simulation work task may be performedin order obtain values various performance measures. In this paper ne~ structural. methodology workstation is presented. Factorial experiments response surface are integrated 111 reduce examined solutions an estimate best...

Most statistical process control (SPC) methods are not suitable for monitoring nonlinear and state-dependent processes. This article introduces the context-based SPC (CSPC) methodology data generated by a finite-memory source. The key idea of CSPC is to monitor attributes comparing two context trees at any period time. first reference tree that represents "in control" behavior process; second monitored tree, periodically from sample sequenced observations, period. Kullback–Leibler (KL)...

Contact mixing plays a key role in the spread of COVID-19. Thus, mobility restrictions varying degrees up to and including nationwide lockdowns have been implemented over 200 countries. To appropriately target timing, location, severity measures intended encourage social distancing at country level, it is essential predict when where outbreaks will occur, how widespread they be. We analyze aggregated, anonymized health data cell phone from Israel. develop predictive models for daily new...

In ordinal classification problems, the class value exhibits a natural order. Usually, these problems are solved as multiclass while discarding ordering form of class. Recently, several research studies have proposed novel methods for aiming to predict predefined objective value. These based on modification splitting criteria decision tree-based algorithms. consider nature data and magnitude potential error from predicted objective. This aims consolidate generalize them any method...

Abstract Current global COVID-19 booster scheduling strategies mainly focus on vaccinating high-risk populations at predetermined intervals. However, these overlook key data: the direct insights into individual immunity levels from active serological testing and indirect information available either through sample-based sero-surveillance, or vital demographic, location, epidemiological factors. Our research, employing an age-, risk-, region-structured mathematical model of disease...

In this paper, we suggest a potential use of data compression measures, such as the Entropy, and Huffman Coding, to assess effects noise factors on reliability tested systems. particular, extend Taguchi method for robust design by computing entropy percent contribution values factors. The new measures are computed already at parameter-design stage, together with traditional S/N ratios enable specification design. Assuming that (some of) should be naturalized, reflects efforts will required...

One highly studied aspect of social networks is the identification influential nodes that can spread ideas in a efficient way. The vast majority works this field have investigated problem identifying set nodes, if "seeded" simultaneously, would maximize information network. Yet, timing aspect, namely, finding not only which should be seeded but also when to seed them, has been sufficiently addressed. In work, we revisit network seeding and demonstrate by simulations how an approach takes...

The immense stream of data from mobile devices during recent years enables one to learn more about human behavior and provide phone users with personalized services. In this work, we identify clusters who share similar mobility behavioral patterns. We analyze trajectories semantic locations find have “lifestyle,” even when they live in different areas. For task, propose a new grouping scheme that is called Lifestyle-Based Clustering (LBC). represent the movement each user by Markov model...

In this article, we evaluate, for the first time, potential of a scheduled seeding strategy influence maximization in real-world setting. We propose methods analyzing historical data to quantify infection probability node with given set properties time and assess infect nodes. Then, examine by large-scale dataset containing both network topology as well nodes' times. Specifically, use proposed demonstrate existence two important effects our dataset: complex contagion effect diminishing...