Abstract
This work presents a method of knowledge discovery in data obtained from Molecular Dynamics Protein Unfolding Simulations. The data under study was obtained from simulations of the unfolding process of the protein Transthyretin (TTR), responsible for amyloid diseases such as Familial Amyloid Polyneuropathy (FAP). Protein unfolding and misfolding are at the source of many amyloidogenic diseases. Thus, the molecular characterization of protein unfolding processes through experimental and simulation methods may be essential in the development of effective treatments. Here, we analyzed the distance variation of each of the 127 amino acids C. (alpha carbon) atoms of TTR to the centre of mass of the protein, along 10 different unfolding simulations - five simulations of WT-TTR and five simulations of L55P-TTR, a highly amyloidogenic TTR variant. Using data mining techniques, and considering all the information of the 10 runs, we identified several clusters of amino acids. For each cluster we selected the representative element and identified events which were used as features. With Association Rules we found patterns that characterize the type of TTR variant under study. These results may help discriminate between amyloidogenic and non-amyloidogenic behaviour among different TTR variants and contribute to the understanding of the molecular mechanisms of FAP.

Abstract

Abstract

Abstract
With so much information available on the web, looking for relevant documents on the Internet has become a difficult task. Temporal features play an important role with the introduction of a time dimension and the possibility to restrict a search by time, recreating a particular moment of a web page set. Despite its importance, temporal information is still under-considered by current search engines, limiting themselves to the capture of the most recent snapshot of the information. In this paper, we describe the architecture of a temporal search engine which uses timelines to browse search results. More specifically, we intend to add a time measure to cluster web page results, by analyzing web page contents, supporting the search of temporal and non-temporal information embedded in web documents.

Abstract

Abstract
Cost sensitive prediction is a key task in many real world applications. Most existing research in this area deals with classification problems. This paper addresses a related regression problem: the prediction of rare extreme values of a continuous variable. These values are often regarded as outliers and removed from posterior analysis. However, for many applications (e.g. in finance, meteorology, biology, etc.) these are the key values that we want to accurately predict. Any learning method obtains models by optimizing some preference criteria. In this paper we propose new evaluation criteria that are more adequate for these applications. We describe a generalization for regression of the concepts of precision and recall often used in classification. Using these new evaluation metrics we are able to focus the evaluation of predictive models on the cases that really matter for these applications. Our experiments indicate the advantages of the use of these new measures when comparing predictive models in the context of our target applications.