Cooperation between expert knowledge and data mining discovered knowledge: Lessons learned

Expert systems are built from knowledge traditionally elicited from the human expert. It is precisely knowledge elicitation from the expert that is the bottleneck in expert system construction. On the other hand, a data mining system, which automatically extracts knowledge, needs expert guidance on the successive decisions to be made in each of the system phases. In this context, expert knowledge and data mining discovered knowledge can cooperate, maximizing their individual capabilities: data mining discovered knowledge can be used as a complementary source of knowledge for the expert system, whereas expert knowledge can be used to guide the data mining process. This article summarizes different examples of systems where there is cooperation between expert knowledge and data mining discovered knowledge and reports our experience of such cooperation gathered from a medical diagnosis project called Intelligent Interpretation of Isokinetics Data, which we developed. From that experience, a series of lessons were learned throughout project development. Some of these lessons are generally applicable and others pertain exclusively to certain project types.

Transformación de series numéricas temporales en secuencias simbólicas

El Trabajo de Fin de Grado aborda el tema del Descubrimiento de Conocimiento en series numéricas temporales, abordando el análisis de las mismas desde el punto de vista de la semántica de las series. La gran mayoría de trabajos realizados hasta la fecha en el campo del análisis de series temporales proponen el análisis numérico de los valores de la serie, lo que permite obtener buenos resultados pero no ofrece la posibilidad de formular las conclusiones de forma que se puedan justificar e interpretar los resultados obtenidos. Por ello, en este trabajo se pretende crear una aplicación que permita realizar el análisis de las series temporales desde un punto de vista cualitativo, en contraposición al tradicional método cuantitativo. De esta forma, quedarán recogidos todos los elementos relevantes de la serie temporal que puedan servir de estudio en un futuro. Para abordar el objetivo propuesto se plantea un mecanismo para extraer de la serie temporal la información que resulta de interés para su análisis. Para poder hacerlo, primero se formaliza el conjunto de comportamientos relevantes del dominio, que serán los símbolos a mostrar en la salida de la aplicación. Así, el método que se ha diseñado e implementado transformará una serie temporal numérica en una secuencia simbólica que recoge toda la semántica de la serie temporal de partida y resulta más intuitiva y fácil de interpretar. Una vez que se dispone de un mecanismo para transformar las series numéricas en secuencias simbólicas, se pueden plantear todas las tareas de análisis sobre dichas secuencias de símbolos. En este trabajo, aunque no se entra en este post-análisis de estas series, sí se plantean distintos campos en los que se puede avanzar en el futuro. Por ejemplo, se podría hacer una medida de la similitud entre dos secuencias simbólicas como punto de partida para la tarea de comparación o la creación de modelos de referencia para análisis posteriores de las series temporales. ---ABSTRACT---This Final-year Project deals with the topic of Knowledge Discovery in numerical time series, addressing time series analysis from the viewpoint of the semantics of the series. Most of the research conducted to date in the field of time series analysis recommends analysing the values of the series numerically. This provides good results but prevents the conclusions from being formulated to allow justification and interpretation of the results. Thus, the purpose of this project is to create an application that allows the analysis of time series, from a qualitative point of view rather than a quantitative one. This way, all the relevant elements of the time series will be gathered for future studies. The design of a mechanism to extract the information that is of interest from the time series is the first step towards achieving the proposed objective. To do this, all the key behaviours in the domain are set, which will be the symbols shown in the output. The designed and implemented method transforms a numerical time series into a symbolic sequence that takes in all the semantics of the original time series and is more intuitive and easier to interpret. Once a mechanism for transforming the numerical series into symbolic sequences is created, the symbolic sequences are ready for analysis. Although this project does not cover a post-analysis of these series, it proposes different fields in which research can be done in the future. For instance, comparing two different sequences to measure the similarities between them, or the creation of reference models for further analysis of time series.

Knowledge-action integration through the media in indigenous communities in Ecuador

The binomial knowledge/action understood under the biunivocal relationship of both components is the basis of planning from a postmodern approach. Within this binomial, social communication gives appropriate information, nurtures the knowledge that leads to transformative action, promotes participation and enhances the community?s self-esteem and recognition; to deeply reflect on action is a source of new knowledge; and communication fosters the adoption of the new knowledge by the community with new actions that feed the process knowledge/action as a planning source. From this approach the project Radio Message is born as a new communication channel with the aim of offering Andean indigenous communities from the area of Cayambe (Ecuador), a series of multidisciplinary training programs that enable transformative action with a strong effect on the life quality in these communities and their importance as social actors. The contents are designed through participatory communication between the training authorities and the communities themselves, analyzing their opportunities and needs. In this research the impact of social media in the development of more than 100 indigenous communities in Cayambe is analyzed.