Automatic blood glucose classification for gestational diabetes with feature selection: decision trees vs neural networks

Automatic blood glucose classification may help specialists to provide a better interpretation of blood glucose data, downloaded directly from patients glucose meter and will contribute in the development of decision support systems for gestational diabetes. This paper presents an automatic blood glucose classifier for gestational diabetes that compares 6 different feature selection methods for two machine learning algorithms: neural networks and decision trees. Three searching algorithms, Greedy, Best First and Genetic, were combined with two different evaluators, CSF and Wrapper, for the feature selection. The study has been made with 6080 blood glucose measurements from 25 patients. Decision trees with a feature set selected with the Wrapper evaluator and the Best first search algorithm obtained the best accuracy: 95.92%.

Presencia en Twitter de los candidatos a las elecciones madrileñas de 2015

Durante los últimos años ha aumentado la presencia de personas pertenecientes al mundo de la política en la red debido a la proliferación de las redes sociales, siendo Twitter la que mayor repercusión mediática tiene en este ámbito. El estudio del comportamiento de los políticos en Twitter y de la acogida que tienen entre los ciudadanos proporciona información muy valiosa a la hora de analizar las campañas electorales. De esta forma, se puede estudiar la repercusión real que tienen sus mensajes en los resultados electorales, así como distinguir aquellos comportamientos que tienen una mayor aceptación por parte de la la ciudadaná. Gracias a los avances desarrollados en el campo de la minería de textos, se poseen las herramientas necesarias para analizar un gran volumen de textos y extraer de ellos información de utilidad. Este proyecto tiene como finalidad recopilar una muestra significativa de mensajes de Twitter pertenecientes a los candidatos de los principales partidos políticos que se presentan a las elecciones autonómicas de Madrid en 2015. Estos mensajes, junto con las respuestas de otros usuarios, se han analizado usando algoritmos de aprendizaje automático y aplicando las técnicas de minería de textos más oportunas. Los resultados obtenidos para cada político se han examinado en profundidad y se han presentado mediante tablas y gráficas para facilitar su comprensión.---ABSTRACT---During the past few years the presence on the Internet of people related with politics has increased, due to the proliferation of social networks. Among all existing social networks, Twitter is the one which has the greatest media impact in this field. Therefore, an analysis of the behaviour of politicians in this social network, along with the response from the citizens, gives us very valuable information when analysing electoral campaigns. This way it is possible to know their messages impact in the election results. Moreover, it can be inferred which behaviours have better acceptance among the citizenship. Thanks to the advances achieved in the text mining field, its tools can be used to analyse a great amount of texts and extract from them useful information. The present project aims to collect a significant sample of Twitter messages from the candidates of the principal political parties for the 2015 autonomic elections in Madrid. These messages, as well as the answers received by the other users, have been analysed using machine learning algorithms and applying the most suitable data mining techniques. The results obtained for each politician have been examined in depth and have been presented using tables and graphs to make its understanding easier.

Evaluating knowledge-based approaches to the multilingual extension of a temporal expression normalizer

The extension to new languages is a well known bottleneck for rule-based systems. Considerable human effort, which typically consists in re-writing from scratch huge amounts of rules, is in fact required to transfer the knowledge available to the system from one language to a new one. Provided sufficient annotated data, machine learning algorithms allow to minimize the costs of such knowledge transfer but, up to date, proved to be ineffective for some specific tasks. Among these, the recognition and normalization of temporal expressions still remains out of their reach. Focusing on this task, and still adhering to the rule-based framework, this paper presents a bunch of experiments on the automatic porting to Italian of a system originally developed for Spanish. Different automatic rule translation strategies are evaluated and discussed, providing a comprehensive overview of the challenge.

Recognition of online handwritten music symbols

Paper submitted to MML 2013, 6th International Workshop on Machine Learning and Music, Prague, September 23, 2013.

Mathematical Morphology: Star/Galaxy Differentiation & Galaxy Morphology Classication

We present an application of Mathematical Morphology (MM) for the classification of astronomical objects, both for star/galaxy differentiation and galaxy morphology classification. We demonstrate that, for CCD images, 99.3 +/- 3.8% of galaxies can be separated from stars using MM, with 19.4 +/- 7.9% of the stars being misclassified. We demonstrate that, for photographic plate images, the number of galaxies correctly separated from the stars can be increased using our MM diffraction spike tool, which allows 51.0 +/- 6.0% of the high-brightness galaxies that are inseparable in current techniques to be correctly classified, with only 1.4 +/- 0.5% of the high-brightness stars contaminating the population. We demonstrate that elliptical (E) and late-type spiral (Sc-Sd) galaxies can be classified using MM with an accuracy of 91.4 +/- 7.8%. It is a method involving fewer 'free parameters' than current techniques, especially automated machine learning algorithms. The limitation of MM galaxy morphology classification based on seeing and distance is also presented. We examine various star/galaxy differentiation and galaxy morphology classification techniques commonly used today, and show that our MM techniques compare very favourably.

Real-Time Gene-Expression Extraction with Fast Classification (FC) Networks

Fast Classification (FC) networks were inspired by a biologically plausible mechanism for short term memory where learning occurs instantaneously. Both weights and the topology for an FC network are mapped directly from the training samples by using a prescriptive training scheme. Only two presentations of the training data are required to train an FC network. Compared with iterative learning algorithms such as Back-propagation (which may require many hundreds of presentations of the training data), the training of FC networks is extremely fast and learning convergence is always guaranteed. Thus FC networks may be suitable for applications where real-time classification is needed. In this paper, the FC networks are applied for the real-time extraction of gene expressions for Chlamydia microarray data. Both the classification performance and learning time of the FC networks are compared with the Multi-Layer Proceptron (MLP) networks and support-vector-machines (SVM) in the same classification task. The FC networks are shown to have extremely fast learning time and comparable classification accuracy.

Exploiting sequence dependencies in the prediction of peroxisomal proteins

Prediction of peroxisomal matrix proteins generally depends on the presence of one of two distinct motifs at the end of the amino acid sequence. PTS1 peroxisomal proteins have a well conserved tripeptide at the C-terminal end. However, the preceding residues in the sequence arguably play a crucial role in targeting the protein to the peroxisome. Previous work in applying machine learning to the prediction of peroxisomal matrix proteins has failed W capitalize on the full extent of these dependencies. We benchmark a range of machine learning algorithms, and show that a classifier - based on the Support Vector Machine - produces more accurate results when dependencies between the conserved motif and the preceding section are exploited. We publish an updated and rigorously curated data set that results in increased prediction accuracy of most tested models.

Precision-recall operating characteristic (P-ROC) curves in imprecise environments

Traditionally, machine learning algorithms have been evaluated in applications where assumptions can be reliably made about class priors and/or misclassification costs. In this paper, we consider the case of imprecise environments, where little may be known about these factors and they may well vary significantly when the system is applied. Specifically, the use of precision-recall analysis is investigated and compared to the more well known performance measures such as error-rate and the receiver operating characteristic (ROC). We argue that while ROC analysis is invariant to variations in class priors, this invariance in fact hides an important factor of the evaluation in imprecise environments. Therefore, we develop a generalised precision-recall analysis methodology in which variation due to prior class probabilities is incorporated into a multi-way analysis of variance (ANOVA). The increased sensitivity and reliability of this approach is demonstrated in a remote sensing application.

Mechatronic system for performing blood pre-transfusion tests

The blood types determination is essential to perform safe blood transfusions. In emergency situations isadministrated the “universal donor” blood type. However, sometimes, this blood type can cause incom-patibilities in the transfusion receptor. A mechatronic prototype was developed to solve this problem.The prototype was built to meet specific goals, incorporating all the necessary components. The obtainedsolution is close to the final system that will be produced later, at industrial scale, as a medical device.The prototype is a portable and low cost device, and can be used in remote locations. A computer appli-cation, previously developed is used to operate with the developed mechatronic prototype, and obtainautomatically test results. It allows image acquisition, processing and analysis, based on Computer Visionalgorithms, Machine Learning algorithms and deterministic algorithms. The Machine Learning algorithmsenable the classification of occurrence, or alack of agglutination in the mixture (blood/reagents), and amore reliable and a safer methodology as test data are stored in a database. The work developed allowsthe administration of a compatible blood type in emergency situations, avoiding the discontinuity of the“universal donor” blood type stocks, and reducing the occurrence of human errors in the transfusion practice.

No free lunch for cross validation

It is known theoretically that an algorithm cannot be good for an arbitrary prior. We show that in practical terms this also applies to the technique of ``cross validation'', which has been widely regarded as defying this general rule. Numerical examples are analysed in detail. Their implications to researches on learning algorithms are discussed.

Probabilistic methods for drug discovery

This thesis introduces a flexible visual data exploration framework which combines advanced projection algorithms from the machine learning domain with visual representation techniques developed in the information visualisation domain to help a user to explore and understand effectively large multi-dimensional datasets. The advantage of such a framework to other techniques currently available to the domain experts is that the user is directly involved in the data mining process and advanced machine learning algorithms are employed for better projection. A hierarchical visualisation model guided by a domain expert allows them to obtain an informed segmentation of the input space. Two other components of this thesis exploit properties of these principled probabilistic projection algorithms to develop a guided mixture of local experts algorithm which provides robust prediction and a model to estimate feature saliency simultaneously with the training of a projection algorithm.Local models are useful since a single global model cannot capture the full variability of a heterogeneous data space such as the chemical space. Probabilistic hierarchical visualisation techniques provide an effective soft segmentation of an input space by a visualisation hierarchy whose leaf nodes represent different regions of the input space. We use this soft segmentation to develop a guided mixture of local experts (GME) algorithm which is appropriate for the heterogeneous datasets found in chemoinformatics problems. Moreover, in this approach the domain experts are more involved in the model development process which is suitable for an intuition and domain knowledge driven task such as drug discovery. We also derive a generative topographic mapping (GTM) based data visualisation approach which estimates feature saliency simultaneously with the training of a visualisation model.

Can a student learn optimally from two different teachers

We explore the effects of over-specificity in learning algorithms by investigating the behavior of a student, suited to learn optimally from a teacher B, learning from a teacher B' ? B. We only considered the supervised, on-line learning scenario with teachers selected from a particular family. We found that, in the general case, the application of the optimal algorithm to the wrong teacher produces a residual generalization error, even if the right teacher is harder. By imposing mild conditions to the learning algorithm form, we obtained an approximation for the residual generalization error. Simulations carried out in finite networks validate the estimate found.

Experiments with two Approaches for Tracking Drifting Concepts

This paper addresses the task of learning classifiers from streams of labelled data. In this case we can face the problem that the underlying concepts can change over time. The paper studies two mechanisms developed for dealing with changing concepts. Both are based on the time window idea. The first one forgets gradually, by assigning to the examples weight that gradually decreases over time. The second one uses a statistical test to detect changes in concept and then optimizes the size of the time window, aiming to maximise the classification accuracy on the new examples. Both methods are general in nature and can be used with any learning algorithm. The objectives of the conducted experiments were to compare the mechanisms and explore whether they can be combined to achieve a synergetic e ect. Results from experiments with three basic learning algorithms (kNN, ID3 and NBC) using four datasets are reported and discussed.

An improved memory management scheme for large scale graph computing engine GraphChi

GraphChi is the first reported disk-based graph engine that can handle billion-scale graphs on a single PC efficiently. GraphChi is able to execute several advanced data mining, graph mining and machine learning algorithms on very large graphs. With the novel technique of parallel sliding windows (PSW) to load subgraph from disk to memory for vertices and edges updating, it can achieve data processing performance close to and even better than those of mainstream distributed graph engines. GraphChi mentioned that its memory is not effectively utilized with large dataset, which leads to suboptimal computation performances. In this paper we are motivated by the concepts of 'pin ' from TurboGraph and 'ghost' from GraphLab to propose a new memory utilization mode for GraphChi, which is called Part-in-memory mode, to improve the GraphChi algorithm performance. The main idea is to pin a fixed part of data inside the memory during the whole computing process. Part-in-memory mode is successfully implemented with only about 40 additional lines of code to the original GraphChi engine. Extensive experiments are performed with large real datasets (including Twitter graph with 1.4 billion edges). The preliminary results show that Part-in-memory mode memory management approach effectively reduces the GraphChi running time by up to 60% in PageRank algorithm. Interestingly it is found that a larger portion of data pinned in memory does not always lead to better performance in the case that the whole dataset cannot be fitted in memory. There exists an optimal portion of data which should be kept in the memory to achieve the best computational performance.