Learning analytics is concerned with collecting, analyzing, and understanding data obtained from students. In this paper, we introduce methods based on data mining and social network analysis to predict a student success. We also focus on identification of different types of data that can be obtained from university information systems. We discuss their influence on the prediction accuracy. We confirm that data about student social behaviour improve accuracy successfully by 3% for one third of the 62 investigated courses.
An information system Xemic applicable in analytical chemistry is described and its use in capillary electrophoresis for searching suitable separation conditions is demonstrated. This system is capable to provide suitable separation conditions even for analytes for which no electrophoretic experiments have been published so far. The system works with a database of components of anionic character the analyses of which have been performed, published in reviewed scientific journals, and included into a database created by an expert. Moreover, the system learned to search also in abstracts or complete scientific articles to find articles dealing with the determination of a substance in a given sample matrix. When no experiments have been published so far for a defined substance in a specific matrix, Xemic shows the separation conditions for determination of the substance regardless of the matrix. When no response can be found for the analyte of interest at all, the system Xemic works like an expert in the field and searches chemically similar substances and offers a series of substances the physicochemical properties of which are close to the followed analyte with respect to the behavior in the electric field, and shows working conditions for their analysis. Thus, the analyst puts only the order in the form of a given analyte in a given matrix and obtains a recommendation of a separation system that should enable to perform a successful separation. The system is not rigid and enables the operator to change the importance of individual attributes used in similarity search so as to obtain a broader or narrower group of similar components. With a certain probability the analyte of interest can be successfully analyzed under separation conditions that suited for the analysis of the most similar substances in the given matrix.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.