Reichenau im Mühlkreis, Austria
Reichenau im Mühlkreis, Austria

Time filter

Source Type

Kogler A.,Data Analysis Systems Group | Traxler P.,Data Analysis Systems Group
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2017

Faults of photovoltaic systems often result in an energy drop and therefore decrease the efficiency of the system. Detecting and analyzing faults is thus an important problem in the analysis of photovoltaic systems data. We consider the problem of estimating the starting time and end time of a fault, i.e. we want to locate the fault in time series data. We assume to know the power output, plane-of-array irradiance and optionally the module temperature. We demonstrate how to use our fault location algorithm to classify shading events. We present results on real data with simulated and real faults. © Springer International Publishing AG 2017.


Kogler A.,Data Analysis Systems Group | Traxler P.,Data Analysis Systems Group
Proceedings - 18th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing, SYNASC 2016 | Year: 2016

We consider the computational problem to learn models from data that are possibly contaminated with outliers. We design and analyze algorithms for robust location and robust linear regression. Such algorithms are essential for solving central problems of robust statistics and outlier detection. We show that our algorithms, which are based on a novel extension of the Median-of-Means method by employing the discrete geometric median, are accurate, efficient and robust against many outliers in the data. The discrete geometric median has many desirable characteristics such as it works for general metric spaces and preserves combinatorial and statistical properties. Furthermore, there is an exact and efficient algorithm to compute it, and an even faster approximation algorithm. We present theoretical and experimental results. In particular, we emphasize the generality of Median-of-Means and its ability to speedup and parallelize algorithms which additionally are accurate and robust against many outliers in the data. © 2016 IEEE.

Loading Data Analysis Systems Group collaborators
Loading Data Analysis Systems Group collaborators