Currently, teaching in higher education is being heavily developed by learning management systems that record the learning behaviour of both students and teachers. The use of learning management systems that include project-based learning and hypermedia resources increases safer learning, and it is proven to be effective in degrees such as nursing. In this study, we worked with 120 students in the third year of nursing degree. Two types of blended learning were applied (more interaction in learning management systems with hypermedia resources vs. none). Supervised learning techniques were applied: linear regression and k-means clustering. The results indicated that the type of blended learning in use predicted 40.4% of student learning outcomes. It also predicted 71.9% of the effective learning behaviors of students in learning management systems. It therefore appears that blended learning applied in Learning Management System (LMS) with hypermedia resources favors greater achievement of effective learning. Likewise, with this type of Blended Learning (BL) a larger number of students were found to belong to the intermediate cluster, suggesting that this environment strengthens better results in a larger number of students. BL with hypermedia resources and project-based learning increase students' learning outcomes and interaction in learning management systems. Future research will be aimed at verifying these results in other nursing degree courses.

No abstract

A natural way of handling imbalanced data is to attempt to equalise the class frequencies and train the classifier of choice on balanced data. For two-class imbalanced problems, the classification success is typically measured by the geometric mean (GM) of the true positive and true negative rates. Here we prove that GM can be improved upon by instance selection, and give the theoretical conditions for such an improvement. We demonstrate that GM is non-monotonic with respect to the number of retained instances, which discourages systematic instance selection. We also show that balancing the distribution frequencies is inferior to a direct maximisation of GM. To verify our theoretical findings, we carried out an experimental study of 12 instance selection methods for imbalanced data, using 66 standard benchmark data sets. The results reveal possible room for new instance selection methods for imbalanced data.

Variscite is an aluminium phosphate mineral widely used as a gemstone in antiquity. Knowledge of the ancient trade in variscite has important implications on the historical appreciation of the commercial and migratory movements of human population. The mining complex of Gavà, which dates from the Neolithic, is one of the oldest underground mine sites in Europe, from where variscite was extracted from several mines and at different depths, providing minerals with different properties and a range of colours. In this work, machine learning algorithms have been used to classify variscite samples from Gavà with regard to the identification of their mine of origin and extraction depth. The final objective of the study was to see if the Raman spectroscopic signatures selected by these algorithms had a key spectral significance related to mineral structure and/or composition and validate the use of these computational procedures as a useful tool for detecting variances in the mineral Raman spectra that could facilitate the assignment of the specimens to each mine.

No abstract