Search results for "ML"
showing 10 items of 1465 documents
Ensemble feature selection with the simple Bayesian classification
2003
Abstract A popular method for creating an accurate classifier from a set of training data is to build several classifiers, and then to combine their predictions. The ensembles of simple Bayesian classifiers have traditionally not been a focus of research. One way to generate an ensemble of accurate and diverse simple Bayesian classifiers is to use different feature subsets generated with the random subspace method. In this case, the ensemble consists of multiple classifiers constructed by randomly selecting feature subsets, that is, classifiers constructed in randomly chosen subspaces. In this paper, we present an algorithm for building ensembles of simple Bayesian classifiers in random sub…
Repeatability Study on a Classifier for Gastric Cancer Detection from Breath Sensor Data
2019
The SNIFFPHONE device is a portable multichannel gas sensor, aiming to detect gastric cancer (GC) from breath samples. It employs gold nanoparticle (GNP) sensors reacting to volatile organic compounds (VOCs) in the exhaled breath, a non-invasive technique to support early diagnosis. This study evaluates the repeatability of the SNIFFPHONE classification result for measurements conducted on healthy subjects over a short period of time of less than 10 minutes. Due to the portable nature of the device, repeatability is studied with respect to varying measurement location. We find the classification results repeatable with a statistically significant 81 % Pearson correlation coefficient, even t…
Multiple Classifiers and Data Fusion for Robust Diagnosis of Gearbox Mixed Faults
2019
Detection and isolation of single and mixed faults in a gearbox are very important to enhance the system reliability, lifetime, and service availability. This paper proposes a hybrid learning algorithm, consisting of multilayer perceptron (MLP)- and convolutional neural network (CNN)-based classifiers, for diagnosis of gearbox mixed faults. Domain knowledge features are required to train the MLP classifier, while the CNN classifier can learn features itself, allowing to reduce the required knowledge features for the counterpart. Vibration data from an experimental setup with gearbox mixed faults is used to validate the effectiveness of the algorithms and compare them with conventional metho…
Facilitating IP deployment in a MARTE-based MDE methodology using IP-XACT: a XILINX EDK case study
2012
International audience; In this paper we present framework for the deployment of hardware IPs at high-levels of abstraction. It is based in a model- driven approach that aims at the automatic generation of Dynamic Partial Reconfiguration designs created in Xilinx Platform Studio (XPS). Contrary to previous approaches, we make use of the IP-XACT standard to facilitate the deployment of hardware IPs, their parameterization and subsequent integration. We propose an extension to the MARTE profile for IP deployment, and we introduce the necessary model transformations to obtain a high- level representation from an IP-XACT component library. These models are then used to create a platform in MART…
Novel Three-Phase Multilevel Inverter With Reduced Components for Low- and High-Voltage Applications
2021
In this article, a novel multilevel topology for three-phase applications, having three-level and hybrid N -level modular configurations, enabling low-, medium-, and high-voltage operations, is presented. The proposed topology has several attractive features, namely reduced component count, being capacitor-, inductor-, and diode-free, lowering cost, control-complexity, and size, and can operate in a wide range of voltages and powers. Selected simulation and experimental results are presented to verify the performance of the proposed topology. Further, the overall efficiency of the topology and loss distribution in switches are studied. Finally, the key features of the proposed topology in t…
Arbiter Meta-Learning with Dynamic Selection of Classifiers and its Experimental Investigation
1999
In data mining, the selection of an appropriate classifier to estimate the value of an unknown attribute for a new instance has an essential impact to the quality of the classification result. Recently promising approaches using parallel and distributed computing have been presented. In this paper, we consider an approach that uses classifiers trained on a number of data subsets in parallel as in the arbiter meta-learning technique. We suggest that information is collected during the learning phase about the performance of the included base classifiers and arbiters and that this information is used during the application phase to select the best classifier dynamically. We evaluate our techn…
Text Classification Using Novel “Anti-Bayesian” Techniques
2015
This paper presents a non-traditional “Anti-Bayesian” solution for the traditional Text Classification (TC) problem. Historically, all the recorded TC schemes work using the fundamental paradigm that once the statistical features are inferred from the syntactic/semantic indicators, the classifiers themselves are the well-established statistical ones. In this paper, we shall demonstrate that by virtue of the skewed distributions of the features, one could advantageously work with information latent in certain “non-central” quantiles (i.e., those distant from the mean) of the distributions. We, indeed, demonstrate that such classifiers exist and are attainable, and show that the design and im…
Infantile Hemangioma Detection using Deep Learning
2020
Infantile hemangiomas are the most common type of benign tumor which appear in the first weeks of life. As currently there is no robust protocol to monitor and assess the hemangioma status, this study proposes a preliminary method to detect the lesion. Therefore, in this paper we describe a hemangiomas classifier based on a linear convolutional neural network architecture. The challenge was to achieve a good classification using a relatively small internal database of 240 images from 40 different patients. The results are promising as the CNN performance evaluation showed a level of accuracy on the test set of 93.84%. Five metrics were calculated to assess the proposed model performances: a…
Modular approach to microswimming
2018
The field of active matter in general and microswimming in particular has experienced a rapid and ongoing expansion over the last decade. A particular interesting aspect is provided by artificial autonomous microswimmers constructed from individual active and inactive functional components into self-propelling complexes. Such modular microswimmers may exhibit directed motion not seen for each individual component. In this review, we focus on the establishment and recent developments in the modular approach to microswimming. We introduce the bound and dynamic prototypes, show mechanisms and types of modular swimming and discuss approaches to control the direction and speed of modular microsw…
A one class classifier for Signal identification: a biological case study
2008
The paper describes an application of a one-class KNN to identify different signal patterns embedded in a noise structured background. The problem become harder whenever only one pattern is well represented in the signal, in such cases one class classifier techniques are more indicated. The classification phase is applied after a preprocessing phase based on a Multi Layer Model (MLM) that provides a preliminary signal segmentation in an interval feature space. The one-class KNN has been tested on synthetic data that simulate microarray data for the identification of nucleosomes and linker regions across DNA. Results have shown a good recognition rate on synthetic data for nucleosome and lin…