Electronically Assisted Astronomy consists in capturing deep sky images with a digital camera coupled to a telescope to display views of celestial objects that would have been invisible through direct observation. This practice generates a large quantity of data, which may then be enhanced with dedicated image editing software after observation sessions. In this study, we show how Image Quality Assessment can be useful for automatically rating astronomical images, and we also develop a dedicated model by using Automated Machine Learning.

In order to analyze the descriptions of various active start-ups, we have developed a web application to retrieve and to analyze available textual data about them. The tool aims at extracting the frequent topics and applying semantic similarity analysis to the start-up descriptions.

Model tree is a useful and convenient method for predictive analytics in data streams, combining the interpretability of decision trees with the efficiency of multiple linear regressions. However, missing values within the data streams is a crucial issue in many real world applications. Often, this issue is solved by pre-processing techniques applied prior to the training phase of the model. In this article we propose a new method that proceeds by estimating and adjusting missing values before the model tree creation. A prototype has been developed and experimental results on several benchmarks show that the method improves the accuracy of the resulting model tree.

Clustering is a popular technique for data mining, knowledge discovery and visual analytics. Unfortunately, clustering interpretation can be uneasy and this difficulty can be overcome by using decision trees to explain cluster assignment. In this work, we propose an evolutionary algorithm to preprocess the data in order to obtain clusters that are simpler to interpret with decision trees. A prototype has been implemented and tested to show the benefits of the approach.

Finding the optimal number of groups in the context of a clustering algorithm is identified as a difficult problem. In this article, we automate this choice for the spectral clustering algorithm with a novel heuristic. Our method is deterministic, and remarkable by its low computational burden. We show its effectiveness with respect to the state of the art, and further investigate assumptions underlying previous work through an empirical study, with the support of synthetic and real data sets.

Decision trees are simple and powerful decision support tools, and their graphical nature can be very useful for visual analysis tasks. However, decision trees tend to be large and hard to display when they are built from complex real world data. This paper proposes an original solution to optimize the visual representation of decision trees obtained from data. The solution combines clustering and feature construction, and introduces a new clustering algorithm that takes into account the visual properties and the accuracy of decision trees. A prototype has been implemented, and the benefits of the proposed method are shown using the results of several experiments performed on the UCI datasets.