In recent years, many research studies are conducted into the use of smart meters data for developping decision-making tools including both analytical, forecasting and display purposes. Forecasting energy generation or consumption demand indeed central problems urban stakeholders (electricity companies planners). These issues helpful to allow them ensuring an efficient planning optimization resources. This paper investigates problem hourly solar irradiance within a Machine Learning (ML)...

The large amount of data collected by smart meters is a valuable resource that can be used to better understand consumer behavior and optimize electricity consumption in cities. This paper presents an unsupervised classification approach for extracting typical patterns from generated electric meters. proposed based on constrained Gaussian mixture model whose parameters vary according the day type (weekday, Saturday or Sunday). methodology applied real dataset Irish households over one year....

Abstract. Nowadays, drinking water utilities need an acute comprehension of the demand on their distribution network, in order to efficiently operate optimization resources, manage billing and propose new customer services. With emergence smart grids, based automated meter reading (AMR), a better understanding consumption modes is now accessible for cities with more granularities. In this context, paper evaluates novel methodology identifying relevant usage profiles from data produced by...

The analysis of time series data issued from smart meters has been studied relatively extensively in the electricity domain. Meanwhile, medium resolution water consumption collected via become possible recently, and research tried to develop statistical machine learning tools order respond different requirements domain, e.g., better understanding behaviors prediction consumption. In present paper, we propose a new predictive approach based on Non-homogeneous Markov Models learn dynamics...

This paper introduces a novel model-based clustering approach for time series which present changes in regime. It consists of mixture polynomial regressions governed by hidden Markov chains. The underlying process each cluster activates successively several regimes during time. parameter estimation is performed the maximum likelihood method through dedicated Expectation-Maximization (EM) algorithm. proposed evaluated using simulated and real-world issued from railway diagnosis application....

Driving errors are considered to be the greatest contributory cause in all road accidents and an important of most fatal accidents. This is particularly case for users powered two-wheeled vehicles (PTWs), perhaps because PTW riders play a greater role control their vehicles' stability than four-wheeled vehicle drivers. Thus, observing analyzing evolution riders' behavior real-life context step identification environment characteristics that constitute risk factor riders. A relevant research...

Mobility demand analysis is increasingly based on smart card data, that are generally aggregated into time series describing the volume of riders along time. These present patterns resulting from multiple external factors. This paper investigates problem decomposing daily ridership data collected at a multimodal transportation hub. The structural models decompose unobserved components. aim decomposition to highlight impact long-term factors, such as trend or seasonality, and exogenous...

A new approach for feature extraction from time series is proposed in this paper. This consists of a specific regression model incorporating discrete hidden logistic process. The parameters are estimated by the maximum likelihood method performed dedicated expectation maximization (EM) algorithm. process, inner loop EM algorithm, using multi-class iterative reweighted least-squares (IRLS) piecewise algorithm and its variant have also been considered comparisons. An experimental study...

Diagnosis of complex systems refers to the problem identifying a breakdown or failure based on an inspection, control test. Monitoring such industrial is essential schedule relevant maintenance actions. We consider automotive subsystem monitor: brake system, because its impact vehicles availability. Through European project [1], data are acquired via in-vehicle communication protocols and additional sensors. This work aims at developing remote diagnostic support tools driven by these data....

This paper is concerned with the issue of online time series segmentation. problem, common in a number applicative fields, continues to receive increasing attention. The present article introduces novel threshold-free sequential segmentation approach. It based on concurrent estimation two models (a model one regressive segment and two-component temporal mixture adapted framework) uses Bayesian Information Criterion decide between models. proposed approach shown be efficient using variety...