Ceit complements its expertise in the technology for extracting useful information from available data with the guidance of professionals in the sector. Ceit has the competences required for data processing proper (database management, feature selection, ETL/ELT processes, data filtering) as well as for applying the appropriate algorithms as a function of a given objective. This includes supervised, unsupervised or semi-supervised algorithms, neural networks and optimization algorithms, in addition to the more traditional machine learning algorithms (SVM, clustering, k-means clustering, logistic regression, random forests, etc.).
In addition to understanding and knowing which algorithm to apply in any given instance, it is also necessary to be able to make decisions when it comes to the concrete implementation of an algorithm, as there are several alternatives and there is no single one that best suits all the different cases. Implementation alternatives include Python libraries, R libraries, Java libraries (WEKA), Mahout, SparkML, FlinkML, etc.