Learning from data streams is a research area of increasing importance. Nowadays, several stream learning algorithms have been developed. Most of them learn decision models that continuously evolve over time, run in resource-aware environments, detect and react to changes in the environment generating data. One important issue, not yet conveniently addressed, is the design of experimental work to evaluate and compare decision models that evolve over time. There are no golden standards for assessing performance in non-stationary environments. This paper proposes a general framework for assessing predictive stream learning algorithms. We defend the use of Predictive Sequential methods for error estimate - the prequential error. The prequential error allows us to monitor the evolution of the performance of models that evolve over time. Nevertheless, it is known to be a pessimistic estimator in comparison to holdout estimates. To obtain more reliable estimators we need some forgetting mechanism. Two viable alternatives are: sliding windows and fading factors. We observe that the prequential error converges to an holdout estimator when estimated over a sliding window or using fading factors. We present illustrative examples of the use of prequential error estimators, using fading factors, for the tasks of: i) assessing performance of a learning algorithm; ii) comparing learning algorithms; iii) hypothesis testing using McNemar test; and iv) change detection using Page-Hinkley test. In these tasks, the prequential error estimated using fading factors provide reliable estimators. In comparison to sliding windows, fading factors are faster and memory-less, a requirement for streaming applications. This paper is a contribution to a discussion in the good-practices on performance assessment when learning dynamic models that evolve over time.

This paper presents and analyzes an incremental system for clustering streaming time series. The Online Divisive-Agglomerative Clustering (ODAC) system continuously maintains a tree-like hierarchy of clusters that evolves with data, using a top-down strategy. The splitting criterion is a correlation-based dissimilarity measure among time series, splitting each node by the farthest pair of streams. The system also uses a merge operator that reaggregates a previously split node in order to react to changes in the correlation structure between time series. The split and merge operators are triggered in response to changes in the diameters of existing clusters, assuming that in stationary environments, expanding the structure leads to a decrease in the diameters of the clusters. The system is designed to process thousands of data streams that flow at a high rate. The main features of the system include update time and memory consumption that do not depend on the number of examples in the stream. Moreover, the time and memory required to process an example decreases whenever the cluster structure expands. Experimental results on artificial and real data assess the processing qualities of the system, suggesting a competitive performance on clustering streaming time series, exploring also its ability to deal with concept drift.

Alzheimer’s disease (AD) has become a problem, owing to its high prevalence in an aging society with no treatment available after onset. However, early diagnosis is essential for preventive intervention to delay disease onset due to its slow progression. The current AD diagnostic methods are typically invasive and expensive, limiting their potential for widespread use. Thus, the development of biomarkers in available biofluids, such as blood, urine, and saliva, which enables low or non-invasive, reasonable, and objective evaluation of AD status, is an urgent task. Here, we reviewed studies that examined biomarker candidates for the early detection of AD. Some of the candidates showed potential biomarkers, but further validation studies are needed. We also reviewed studies for non-invasive biomarkers of AD. Given the complexity of the AD continuum, multiple biomarkers with machine-learning-classification methods have been recently used to enhance diagnostic accuracy and characterize individual AD phenotypes. Artificial intelligence and new body fluid-based biomarkers, in combination with other risk factors, will provide a novel solution that may revolutionize the early diagnosis of AD.

Introduction: Colectomy is the major event which may significantly affect the outcome of ulcerative colitis both in terms of quality of life and mortality. This paper aims to identify clinical prognostic factors that may be significantly associated with this event. Methods: Pubmed, ISI Web of Knowledge and Scopus were searched in order to identify studies investigating the association between clinical factors in adult patients with UC and studied events. The clinical factors evaluated in this meta-analysis were: gender, smoking habits, localization of disease, use of corticosteroids and episodes of hospitalization. Results: Of the 3753 initially selected papers, twenty were included. The analysis showed a significantly lower risk of colectomy for female patients (OR=0.78 [0.68;0.90])) and for smoking patients (OR=0.55 [0.33;0.91]), and a higher risk for patients with extensive disease (OR=3.68 [2.39;5.69]), for patients who took corticosteroids at least once (OR=2.10 [1.05;4.22]) and for patients who were hospitalized (OR=4.13 [3.23;5.27]). Conclusion: Gender, smoking habits, localization of disease, need for corticosteroids and hospitalization were all significantly associated with UC prognosis. These results could clarify the relative influence of these and other prognostic factors in the natural course of the disease and, therefore, help a better management approach, thus improving the follow up of patients.