This paper identifies, by means of video and Kinect data, a set of predictors that estimate the presentation skills of 448 individual students. Two evaluation criteria were predicted: eye contact and posture and body language. Machine-learning evaluations resulted in models that predicted the performance level (good or poor) of the presenters with 68% and 63% of correctly classified instances, for eye contact and postures and body language criteria, respectively. Furthermore, the results suggest that certain features, such as arms movement and smoothness, provide high significance on predicting the level of development for presentation skills. The paper finishes with conclusions and related ideas for future work.
Los servicios de las empresas de desarrollo de software se basan en producir productos de software de alta calidad. La calidad de los productos de software se asegura mediante la aplicación de prácticas de ingeniería de software a lo largo de todo el proceso de desarrollo. Para mejorar estos procesos, es necesario adaptar los modelos de proceso de mejora de software a las empresas de acuerdo con sus propias características. Este artículo ofrece un análisis de la aplicación de diversos modelos de mejora de procesos de software en pequeñas y medianas empresas. Finalmente, los resultados aquí presentados evidencian la influencia de los factores inherentes de las empresas y de sus equipos de trabajo en la elección de un modelo de proceso de mejora de software específico. AbstractThe services of the software development companies are based on producing high quality software products. Software products quality is ensured by applying software engineering practices throughout the development process. In order to improve these processes, it is necessary adapting the software improvement process models into companies according to their own characteristics. This paper offers an analysis of the application of various software process improvement models in small and medium size enterprises. Finally, results presented here show the influence of the inherent factors of companies and their work teams into choosing a specific software improvement process model.
Cities are complex systems evolving constantly. Thus, it is necessary to improve the way we collect intra-urban data in order to quantify such evolution. We propose a methodology to transform geo-located tweets into labels for different areas of a given city using DBPedia, Wikipedia and Foursquare categories. We conduct experiments using 77K geolocated tweets posted in Milan during November and December 2013 and feed a clustering algorithm with the annotated tweets to produce dynamic thematic maps. Since, Twitter is the most popular platform for publishing short public messages, to generate crowd-sourced city maps. The results suggest that we can accurately find different functional areas on different temporal bands.