Visualization Based Data Mining for Comparison Between Two Solar Cell Libraries

Abstract

Material informatics may provide meaningful insights and powerful predictions for the development of new and efficient Metal Oxide (MO) based solar cells. The main objective of this paper is to establish the usefulness of data reduction and visualization methods for analyzing data sets emerging from multiple all-MOs solar cell libraries. For this purpose, two libraries, TiO |Co O and TiO |Co O |MoO , differing only by the presence of a MoO layer in the latter were analyzed with Principal Component Analysis and Self-Organizing Maps. Both analyses suggest that the addition of the MoO layer to the TiO |Co O library has affected the overall photovoltaic (PV) activity profile of the solar cells making the two libraries clearly distinguishable from one another. Furthermore, while MoO had an overall favorable effect on PV parameters, a sub-population of cells was identified which were either indifferent to its presence or even demonstrated a reduction in several parameters.

RANdom SAmple Consensus (RANSAC) algorithm for material-informatics: application to photovoltaic solar cells

“…This conclusion was further supported by a computational analysis [36]. Figure 8 shows that RANSAC’s prediction is in line with this proposition (i.e., to achieve relatively high IQE values, the thickness of the layer must be low, smaller than 150 nm and this property is also influenced by the thickness of the layer).…”

“…RANSAC is a modeling tool widely used in the Image Processing field [34–36] primarily for image noise filtration. The algorithm produces and validates a linear QSAR model based on the Minimum Least Square (LMS) method by (1) filtering noisy samples (i.e., outliers), (2) selecting the best features (i.e., descriptors), (3) deriving a QSAR model from training set samples and (4) predicting the activity of test set samples while invoking the concept of applicability domain, all in a single process without the need of complementary processes.…”

“…The two libraries comprised of based solar cells were previously modeled using k nearest neighbors ( kNN ) and genetic algorithm allowing for a facile comparison between the performances of the different algorithms. The third library ( was previously analyzed using visualization methods [36]. We demonstrate that the RANSAC algorithm filters the sample space from noisy data (i.e., outliers), automatically selects descriptors previous shown to correlate with key PV properties and generates models with good predictive statistics for these properties.…”

Materials Informatics: Statistical Modeling in Material Science

“…Yet the method can also be used to detect interesting phenomena in solar cell libraries. In one study Yosipof et al . have compared two libraries, TiO 2 |Co 3 O 4 and TiO 2 |Co 3 O 4 |MoO 3 , which differ only by the presence of the MoO 3 layer in the latter.…”

Property Prediction of Organic Donor Molecules for Photovoltaic Applications Using Extremely Randomized Trees

“…The screening of OPV materials could be semi‐automated through utilization of various modeling techniques (finite element to ab initio and molecular modeling). Yosipof et al . establishes the importance of data reduction and visualization using Principle Component Analysis and Self Organizing Maps, wherein two metal oxide solar cell libraries are analyzed.…”