Time Series Complexities and Their Relationship to Forecasting Performance.

Entropy is a key concept in the characterization of uncertainty for any given signal, and its extensions such as Spectral Entropy and Permutation Entropy. They have been used to measure the complexity of time series. However, these measures are subject to the discretization employed to study the states of the system, and identifying the relationship between complexity measures and the expected performance of the four selected forecasting methods that participate in the M4 Competition. This relationship allows the decision, in advance, of which algorithm is adequate. Therefore, in this paper, we found the relationships between entropy-based complexity framework and the forecasting error of four selected methods (Smyl, Theta, ARIMA, and ETS). Moreover, we present a framework extension based on the Emergence, Self-Organization, and Complexity paradigm. The experimentation with both synthetic and M4 Competition time series show that the feature space induced by complexities, visually constrains the forecasting method performance to specific regions; where the logarithm of its metric error is poorer, the Complexity based on the emergence and self-organization is maximal.

Melting Temperature Estimation of Imidazole Ionic Liquids with Clustering Methods.

Ionic liquids (ILs) are ionic compounds with low melting points that can be designed to be used in an extensive set of commercial and industrial applications. However, the design of ILs is limited by the quantity and quality of the available data in the literature; therefore, the estimation of physicochemical properties of ILs by computational methods is a promising way of solving this problem, since it provides approximations of the real values, resulting in savings in both time and money. We studied two data sets of 281 and 134 liquids based on the molecule imidazole that were analyzed with QSPR techniques. This paper presents a software architecture that uses clustering techniques to improve the robustness of estimation models of the melting point of ILs. These results indicate an error of 6.25% in the previously unmodeled data set and an error of 4.43% in the second data set. We have an improvement with the second data set of 1.81% over the last results previously found.

A Predictive Model for Guillain-Barré Syndrome Based on Single Learning Algorithms.

Background. Guillain-Barré Syndrome (GBS) is a potentially fatal autoimmune neurological disorder. The severity varies among the four main subtypes, named as Acute Inflammatory Demyelinating Polyneuropathy (AIDP), Acute Motor Axonal Neuropathy (AMAN), Acute Motor Sensory Axonal Neuropathy (AMSAN), and Miller-Fisher Syndrome (MF). A proper subtype identification may help to promptly carry out adequate treatment in patients. Method. We perform experiments with 15 single classifiers in two scenarios: four subtypes' classification and One versus All (OvA) classification. We used a dataset with the 16 relevant features identified in a previous phase. Performance evaluation is made by 10-fold cross validation (10-FCV). Typical classification performance measures are used. A statistical test is conducted in order to identify the top five classifiers for each case. Results. In four GBS subtypes' classification, half of the classifiers investigated in this study obtained an average accuracy above 0.90. In OvA classification, the two subtypes with the largest number of instances resulted in the best classification results. Conclusions. This study represents a comprehensive effort on creating a predictive model for Guillain-Barré Syndrome subtypes. Also, the analysis performed in this work provides insight about the best single classifiers for each classification case.

Multiphase Simulated Annealing Based on Boltzmann and Bose-Einstein Distribution Applied to Protein Folding Problem.

A new hybrid Multiphase Simulated Annealing Algorithm using Boltzmann and Bose-Einstein distributions (MPSABBE) is proposed. MPSABBE was designed for solving the Protein Folding Problem (PFP) instances. This new approach has four phases: (i) Multiquenching Phase (MQP), (ii) Boltzmann Annealing Phase (BAP), (iii) Bose-Einstein Annealing Phase (BEAP), and (iv) Dynamical Equilibrium Phase (DEP). BAP and BEAP are simulated annealing searching procedures based on Boltzmann and Bose-Einstein distributions, respectively. DEP is also a simulated annealing search procedure, which is applied at the final temperature of the fourth phase, which can be seen as a second Bose-Einstein phase. MQP is a search process that ranges from extremely high to high temperatures, applying a very fast cooling process, and is not very restrictive to accept new solutions. However, BAP and BEAP range from high to low and from low to very low temperatures, respectively. They are more restrictive for accepting new solutions. DEP uses a particular heuristic to detect the stochastic equilibrium by applying a least squares method during its execution. MPSABBE parameters are tuned with an analytical method, which considers the maximal and minimal deterioration of problem instances. MPSABBE was tested with several instances of PFP, showing that the use of both distributions is better than using only the Boltzmann distribution on the classical SA.

EMRlog method for computer security for electronic medical records with logic and data mining.

The proper functioning of a hospital computer system is an arduous work for managers and staff. However, inconsistent policies are frequent and can produce enormous problems, such as stolen information, frequent failures, and loss of the entire or part of the hospital data. This paper presents a new method named EMRlog for computer security systems in hospitals. EMRlog is focused on two kinds of security policies: directive and implemented policies. Security policies are applied to computer systems that handle huge amounts of information such as databases, applications, and medical records. Firstly, a syntactic verification step is applied by using predicate logic. Then data mining techniques are used to detect which security policies have really been implemented by the computer systems staff. Subsequently, consistency is verified in both kinds of policies; in addition these subsets are contrasted and validated. This is performed by an automatic theorem prover. Thus, many kinds of vulnerabilities can be removed for achieving a safer computer system.

Towards a predictive model for Guillain-Barré syndrome.

The severity of Guillain-Barré Syndrome (GBS) varies among subtypes, which can be mainly Acute Inflammatory Demyelinating Polyneuropathy (AIDP), Acute Motor Axonal Neuropathy (AMAN), Acute Motor Sensory Axonal Neuropathy (AMSAN) and Miller-Fisher Syndrome (MF). In this study, we use a real dataset that contains clinical, serological, and nerve conduction tests data obtained from 129 GBS patients. We apply C4.5 decision tree, SVM (Support Vector Machines) using a Gaussian kernel, and kNN (k Nearest Neighbour) to predict four GBS subtypes. Accuracies were calculated and averaged across 30 10-fold cross-validation (10-FCV) runs. C4.5 obtained 0.9211 (±0.0109), kNN 0.9179 (±0.0041), and SVM 0.9154 (±0.0069). This is an ongoing research project and further experiments are being conducted.

Feature selection for better identification of subtypes of Guillain-Barré syndrome.

Guillain-Barré syndrome (GBS) is a neurological disorder which has not been explored using clustering algorithms. Clustering algorithms perform more efficiently when they work only with relevant features. In this work, we applied correlation-based feature selection (CFS), chi-squared, information gain, symmetrical uncertainty, and consistency filter methods to select the most relevant features from a 156-feature real dataset. This dataset contains clinical, serological, and nerve conduction tests data obtained from GBS patients. The most relevant feature subsets, determined with each filter method, were used to identify four subtypes of GBS present in the dataset. We used partitions around medoids (PAM) clustering algorithm to form four clusters, corresponding to the GBS subtypes. We applied the purity of each cluster as evaluation measure. After experimentation, symmetrical uncertainty and information gain determined a feature subset of seven variables. These variables conformed as a dataset were used as input to PAM and reached a purity of 0.7984. This result leads to a first characterization of this syndrome using computational techniques.

Attribute selection impact on linear and nonlinear regression models for crop yield prediction.

Efficient cropping requires yield estimation for each involved crop, where data-driven models are commonly applied. In recent years, some data-driven modeling technique comparisons have been made, looking for the best model to yield prediction. However, attributes are usually selected based on expertise assessment or in dimensionality reduction algorithms. A fairer comparison should include the best subset of features for each regression technique; an evaluation including several crops is preferred. This paper evaluates the most common data-driven modeling techniques applied to yield prediction, using a complete method to define the best attribute subset for each model. Multiple linear regression, stepwise linear regression, M5' regression trees, and artificial neural networks (ANN) were ranked. The models were built using real data of eight crops sowed in an irrigation module of Mexico. To validate the models, three accuracy metrics were used: the root relative square error (RRSE), relative mean absolute error (RMAE), and correlation factor (R). The results show that ANNs are more consistent in the best attribute subset composition between the learning and the training stages, obtaining the lowest average RRSE (86.04%), lowest average RMAE (8.75%), and the highest average correlation factor (0.63).

Chaotic multiquenching annealing applied to the protein folding problem.

The Chaotic Multiquenching Annealing algorithm (CMQA) is proposed. CMQA is a new algorithm, which is applied to protein folding problem (PFP). This algorithm is divided into three phases: (i) multiquenching phase (MQP), (ii) annealing phase (AP), and (iii) dynamical equilibrium phase (DEP). MQP enforces several stages of quick quenching processes that include chaotic functions. The chaotic functions can increase the exploration potential of solutions space of PFP. AP phase implements a simulated annealing algorithm (SA) with an exponential cooling function. MQP and AP are delimited by different ranges of temperatures; MQP is applied for a range of temperatures which goes from extremely high values to very high values; AP searches for solutions in a range of temperatures from high values to extremely low values. DEP phase finds the equilibrium in a dynamic way by applying least squares method. CMQA is tested with several instances of PFP.