A method is proposed for testing whether the failure of the first specimen in a test machine that contains multiple specimens shortens the time to failure of the other specimens in that unit. The time to failure is modeled as a two-parameter Weibull distribution. Two estimates of the shape parameter are constructed: an estimate based on the first failures in each of k identical test units, and an estimate based on the ratios of the second to the first failure times within each unit. The supposition is that these ratios will be shortened if the first failure has influenced the second, causing the estimate of the shape parameter constructed from these ratios to tend to be larger than the estimate obtained from the set of first failures. The null distribution of the ratio of the two shape parameter estimates was determined by Monte Carlo sampling for the case of k units containing exactly two specimens each with k = 5(1)20. The power is determined under the assumption that the first failure causes the remain...

Testing for dependency of failure times in life testing

Many prediction, decision-making, and control architectures rely on online learned Gaussian process (GP) models. However, most existing GP regression algorithms assume a single generative model, leading to poor predictive performance when the data are nonstationary, i.e., generated from multiple switching processes. Furthermore, existing methods for GP regression over nonstationary data require significant computation, do not come with provable guarantees on correctness and speed, and many only work in batch settings, making them ill-suited for real-time prediction. We present an efficient online GP framework, GP-non-Bayesian clustering (GP-NBC), which addresses these computational and theoretical issues, allowing for real-time changepoint detection and regression using GPs. Our empirical results on two real-world data sets and two synthetic data set show that GP-NBC outperforms state-of-the-art methods for nonstationary regression in terms of both regression error and computation. For example, it outperforms Dirichlet process GP clustering with Gibbs sampling by 98% in computation time reduction while the mean absolute error is comparable.

Online Regression for Data With Changepoints Using Gaussian Processes and Reusable Models

Recently, there has been an increasing interest in offloading the 3GPP LTE data by using device-to-device (D2D) communications between devices. However, the peer discovering is challenging, especially in the indoor environment, since traditional users use a cellular signal to find peers, leading to incurring interference to other cellular users. In this paper, we propose ROOMMATEs, a novel approach for indoor peer discovery process, which is the enabler for indoor D2D communications in long term evolution (LTE) networks. It is a centralized approach utilizing, but not limited to, the ubiquitous WiFi network/femtocell network, combining with eNodeB in order to deliver the best results. ROOMMATEs is an unsupervised, yet energy efficient algorithm that can find surrounding user equipments (UEs) while minimizing interference and consuming much less energy. Based on the results, ROOMMATEs proves to be highly energy efficient, saving on average 24% per UE and improving signal-to-interference-plus-noise ratio in order of tens of decibels compared to other approaches. Moreover, ROOMMATEs provides indoor place identification for UEs with high accuracy using a small number of observations. ROOMMATEs is robust to missing and noisy data, works with various UEs’ brands/models, and requires no user interactions.

ROOMMATEs: An Unsupervised Indoor Peer Discovery Approach for LTE D2D Communications

Demographics and self-reported well-being of Brazilian adults as a function of pet ownership: A pilot study

https://dyuthi.cusat.ac.in/xmlui/bitstream/handle/purl/2369/Dyuthi-T0641.pdf?sequence=1

A bivariate pareto distribution with freund's dependence structure

Nonparametric Statistics: Theory and Methods

Nonparametric bayesian methods

Life times of load sharing parallel systems have been considered in thestatistical literature at least since Daniels (1945). The main characteristicof such a two component system is that after the failure of one componentthe surviving component has to shoulder extra load and hence is prone tofailure at an earlier time than what is expected under the original model. Inother situations, the failure of one component may release extra resourcesto the other, thus delaying the system failure. Gross et al (1971) observedthat similar considerations aﬀect the functioning of a two organ system. Inthis paper we ﬁrst consider several observations schemes and identiﬁabilityissues under them. Then we construct a general semiparametric bivariatefamily of distributions which explicitly models this phenomenon throughproportional conditional hazards. McCool (2006) has suggested a test forthe hypothesis that the failures take place according to the original modelagainst the alternative hypothesis that the second failure takes place earlierthan warranted within the Weibull model. We propose nonparametric testsfor the same problem which may be used for any continuous distributionfor the component life times. We obtain estimates of the power of the testand observe that it is quite high even for moderately distant alternatives.The tests are applied to several real data sets to illustrate their use.Key words: Bivariate distributions, censoring, conditional distribution,early failures, order statistics, proportional hazards, semiparametric family,sharing resources.

/pdf/two-component-load-sharing-systems-with-applications-to-2eng30och6.pdf

Two component load sharing systems with applications to Biology

This chapter describes the background material related to the nonparametric density estimation. Techniques such as histograms (together with its extension, known as ASH, see Sect. 2.3), Parzen windows and k-nearest neighbors are at the core of the applications of nonparametric density estimation. For that reason, we decided to include a chapter describing these for the sake of completeness and to allow less experienced readers develop their intuitions in terms of the nonparametric estimation. Most of the material is presented taking into account only the univariate case; extending the results to cover more than one variable, however, is often a straightforward task. The chapter is organized as follows: Sect. 2.2 presents a short overview of the fundamental concepts related to histograms. Section2.3 is devoted to a description of a smart extension of certain well-known histograms aimed at avoiding some of their drawbacks. Section2.4 presents basic concepts related to the nonparametric density estimation. Section2.5 is devoted to the Parzen windows, while Sect. 2.6 to the k-nearest neighbors approach.

Nonparametric density estimation

Competing risks data arise when the study units are exposed to several risks at the same time but it is assumed that the eventual failure of a unit is due to only one of these risks, which is called the “cause of failure”. Statistical inference procedures when the time to failure and the cause of failure are observed for each unit are well documented. In some applications, it is possible that the cause of failure is either missing or masked for some units. In this article, we review some statistical inference procedures used when the cause of failure is missing or masked for some units.

Uttara Naik-Nimbalkar

Papers

Nonparametric Statistics: Theory and Methods

Nonparametric bayesian methods

Two component load sharing systems with applications to Biology

Nonparametric density estimation

On Competing Risks with Masked Failures