Data Envelopment Analysis (DEA) is a nonparametric technique which measures the inefficiency of Decision Making Unit (DMUs), where multiple input and output variables exist. A DMU is considered efficient if, within an array of production possibilities, there is not an activity that with less inputs produces more outputs. However, for inefficient DMUs, the DEA provides a score that allows the analysis of inputs and output that need to be improved to be efficient.
Application areas that the MODELEM group has worked in: Education, Economics and Business, Environmental science 

The time series wavelet analysis allows the analysis and the decompose of signs in various frequencies o scales over time, exceeding the limitations of the Fourier analysis. There is a wide range of applications for the analysis, from processing digital images to medicine, physics and engineering. In fact, the group MODELEM has applied the wavelet analysis to the comparison between economic time series and the degree quantification of aperiodic signals.
Application areas that de group MODELEM has worked in: Economics and Business, Dynamic Systems. 

One of the basic principles of the efficient management of any process, product, service, or institution is the approach towards decision-making, which must be made based on the information available. To obtain relevant information for this decision-making it is required to have a data set, whose analysis must provide summarised or synthetic methods that allow to understand the studied phenomenon. This is the place where indicators are defined, among many other meanings, as the simplest way to reduce the great amount of data keeping the essential information for the issues posed to the data.
Indicators are not usually managed individually, but as part of structured systems of indicators that allow to measure the level of achievement of the objectives that any organisation suggests. 

Moreover, when the objective is to measure multidimensional concepts that can’t be captured by simple indicators, composite indicators are built. Those indicators are formed when simple indicators (individuals) that compile in only one index, based on an underlying model of the multidimensional concept being measured. The MODELEM groupis specialised in the construction of the different stages required to define correctly composite indicators.

Application areas the MODELEM group has worked on: Education, Economy, and Business.

The multi-criteria analysis is a methodology that allows to help to make decision when there are different alternatives that must be ordered following a series of criteria, which, in many cases, are contradictory or conflict. Among the different methods available, the MODELEM group is specialised in distance-based methods, such as TOPSIS or VIKOR, both in its standard versions and in its unweighted ones (in which the decision-maker is not required to define the weight or relative importance of the different criteria). Additionally, they have also analysed cases in which there is uncertainty with the data, caused by the intrinsic lack of definition of the criteria (fuzzy methods).

Application areas the MODELEM group has worked on: Education, Employment policy, Algorithm analysis.

Machine learning is a branch of artificial intelligence, interferential statistics, and computing, in which algorithms for pattern detection are developed. There is usually a distinction between supervised learning, in which data is labelled and the algorithms have to either predict labels or classify them, and unsupervised learning, in which data is not labelled and the algorithms try to define structures in the data set. Within the first type, decision trees, neuronal networks, and support vector machines stand out, whereas within the second type, the most used algorithms are the clustering ones and those of dimensionality reduction (principal component analysis (PCA)). The MODELEM group has used machine learning techniques to design neuronal prediction networks in temporal series, clustering techniques in temporal series using distances based on wavelets.

Application areas the MODELEM group has worked on: econometrics, finances, water consumption management in urban areas, sales prediction.