Papers by Ricardo Rodriguez Jorge
Arrhythmia disease classification using a higher-order neural unit
2015 Fourth International Conference on Future Generation Communication Technology (FGCT), 2015
This paper presents a quadratic neural unit with error backpropagation learning algorithm to clas... more This paper presents a quadratic neural unit with error backpropagation learning algorithm to classify electrocardiogram arrhythmia disease. The electrocardiogram arrhythmia classification scheme consists of data acquisition, feature extraction, feature reduction, and a quadratic neural unit classifier to discriminate three different types of arrhythmia. A total of 44 records were obtained from MIT-BIH arrhythmia database to test the efficiency of arrhythmia disease classification method, the obtained results were a specificity of 97.60 % and a sensitivity of 97.05 %. The best accuracy classification rate obtained using the presented approach has been of 98.16 %.
MENDEL, 2019
Complex systems are very hard to describe by some unified language and calculus. In cases when th... more Complex systems are very hard to describe by some unified language and calculus. In cases when their nature is very heterogeneous is possible to use with advantage state description. Formalization of operations on the set of states usually leads to partial algebras. The work with partial algebras is rather difficult and unpractical. From this reason some methods approximating partial algebras by some more symmetrical objects are searched for. In this paper there is proposed an approximation of this algebras by free cyclic groups. Then using the combination of Matroid Theory and Ramsey theory of graph the prediction of a possible appearance of emergent situation is executed. Data and knowledge used in the paper for the demonstration of developed method application are from the field of Ecology.
A grey-box model for real-time control and monitoring
Advances on P2P, Parallel, Grid, Cloud and Internet Computing. 3PGCIC 2024. Lecture Notes on Data Engineering and Communications Technologies, 2024
Adaptive Analysis of Electrocardiogram Prediction Using a Dynamic Cubic Neural Unit
Lecture notes in networks and systems, 2022
n this work, the implementation of a dynamic cubic neural unit for the prediction of heartbeats u... more n this work, the implementation of a dynamic cubic neural unit for the prediction of heartbeats using a wireless method is presented. The data were recorded with the BITalino biomedical acquisition card using its ECG input and output module via Bluetooth. This paper aims to predict a prediction horizon according to the learning rate, the number of samples used to train the model, and the specified times required for training. The signal (input) was acquired from electrodes, which were placed on the surface of the chest near the heart. The signal was visualized and presented through a graphical interface. For the interface evaluation, tests are performed using the obtained signal in real time.
A Modified Version of K-Means Algorithm
In this work is presented a modified version of the K-Means which identifies cluster stability. T... more In this work is presented a modified version of the K-Means which identifies cluster stability. The stability is defined by a threshold based on a percentage of the largest displacement of centroid at first iteration. A cluster is considered stable when the largest centroid displacement in the current iteration achieves the 10% of threshold, and objects that remains in the same cluster in two consecutive iterations are removed from the classification phase in subsequent iterations. Eight different instances were used to validate the proposal, three synthetics and five reals. The modified version was compared against the standard and three related work versions. Results shows that the proposal reduced the execution time up to 92.14% regarding the standard version with only a 3.73% in the quality reduction. Despite the new version do not has the major reduction time in all cases, the algorithm reaches the best values for quality of grouping.
Quadratic neural unit is a good compromise between linear models and neural networks for industrial applications
The paper discusses the quadratic neural unit (QNU) and highlights its attractiveness for industr... more The paper discusses the quadratic neural unit (QNU) and highlights its attractiveness for industrial applications such as for plant modeling, control, and time series prediction. Linear systems are still often preferred in industrial control applications for their solvable and single solution nature and for the clarity to the most application engineers. Artificial neural networks are powerful cognitive nonlinear tools, but their nonlinear strength is naturally repaid with the local minima problem, overfitting, and high demands for application-correct neural architecture and optimization technique that often require skilled users. The QNU is the important midpoint between linear systems and highly nonlinear neural networks because the QNU is relatively very strong in nonlinear approximation; however, its optimization and performance have fast and convex-like nature, and its mathematical structure and the derivation of the learning rules is very comprehensible and efficient for implementation.
IEEE LATIN AMERICA TRANSACTIONS, 2020
Scene classification is a computer vision task that aims to identify the kind of scene (such as f... more Scene classification is a computer vision task that aims to identify the kind of scene (such as forest, mountain, beach) where a picture was taken. Scene classification has application in the development of automatic surveillance systems, robotic navigation, content-based image retrieval systems among other areas. According to how a scene is recognized, scene classification algorithms can be divided in two categories: based on object recognition and based on low-level image features obtained by applying descriptors. This paper proposes a new binary descriptor called Local Lineal Binary Pattern and a new framework that allows the combination of the new binary descriptor with another
local and global descriptors in order to improve the automatic classification of scene images. This new binary descriptor increases the spatial support for its calculation allowing to add more spatial information than the traditional binary descriptors such as the Local Binary Pattern and the Modified Census Transform. Experiments conducted over indoor and outdoor scene datasets show that the new proposed descriptor and framework help to improve the results obtained by related works.
Internet of things-assisted architecture for QRS complex detection in real time
Internet of Things
Abstract This paper presents the development of an Internet of Things-assisted architecture for Q... more Abstract This paper presents the development of an Internet of Things-assisted architecture for QRS complex detection in an electrocardiogram regardless of the age and physiological characteristics of the patient. Detection of the QRS complex is affected by the abnormalities and quality in electrocardiogram recordings; the proposed method can detect QRS complex despite these challenges. Electrocardiogram continuous signal acquisition is performed with the BITalino biomedical data acquisition card. Electrocardiogram signals typically suffer from (a) premature atrial complexes, (b) premature ventricular complexes, (c) low signal-to-noise ratio, (d) right bundle branch blocks, (e) left bundle branch blocks, and (f) non-stationary effects. Interestingly, the signal processing is implemented by means of a bandpass filter, followed by a numerical derivative. Next, the Hilbert transform and the adaptive threshold technique are implemented to detect the QRS complex. Tests are performed to evaluate the Internet of Things-assisted architecture using the obtained signal in real time. Results, and the simplicity of the architecture, demonstrate that it is suitable for wearable, portable, and battery-operated electrocardiogram acquisition card.
Non Linear Fitting Methods for Machine Learning
This manuscript presents an analysis of numerical fitting methods used for solving classification... more This manuscript presents an analysis of numerical fitting methods used for solving classification problems as discriminant functions in machine learning. Non linear polynomial, exponential, and trigonometric models are mathematically deduced and discussed. Analysis about their pros and cons, and their mathematical modelling are made on what method to chose for what type of highly non linear multi-dimension problems are more suitable to be solved. In this study only deterministic models with analytic solutions are involved, or parameters calculation by numeric methods, which the complete model can subsequently be treated as a theoretical model. Models deduction are summarised and presented as a survey.
Advances on P2P, Parallel, Grid, Cloud and Internet Computing, Lecture Notes on Data Engineering and Communications Technologies 13, 2017
In this paper, the combination of the Hilbert-Huang Transform, fuzzy logic and an embedding theor... more In this paper, the combination of the Hilbert-Huang Transform, fuzzy logic and an embedding theorem is described to predict the short-term exchange rate from United States dollar to Czech Koruna. By Using the Hilbert-Huang Transform as an adaptive filter, the proposed method decreases the embedding dimension space from five (original samples) to four (de-noising samples). This dimension space provides the number of inputs to the fuzzy rule base system, which causes the number of rules, the time for training and the inference process to decrease. Experimental results indicated that this method achieves higher accuracy prediction than the direct use of original data.
Artificial Neural Networks: Challenges in Science and Engineering Applications
In this paper, artificial neural networks applications in the prediction field are described. The... more In this paper, artificial neural networks applications in the prediction field are described. The aim is to analyze the potentialities of conventional neural networks, such as feedforward neural networks, recurrent neural networks; and also, the potentialities of nonconventional neural networks composed typically by higher-order neural units. Finally, experimental analysis of long-term prediction of non-stationary time series (Mackey-Glass) is presented as well. The resulting prediction made by the proposed neural models feedforward multilayer perceptron and quadratic neural unit show high prediction accuracy for non-stationary time series.
Fuzzy Control Proposal for the Climate of a Homemade Greenhouse
In this study, we develop a fuzzy control that has the ability to reduce energy use and uncertain... more In this study, we develop a fuzzy control that has the ability to reduce energy use and uncertainties in crop production by reducing or increasing the temperature in a homemade urban greenhouse by opening a window to two different angles to allow the temperature inside the greenhouse to be equal to the outside temperature. We use hydroponic tomato cultivation as our test case because tomatoes are part of the goods and services category for the “Índice Nacional de Precios al Consumidor” in México, which is an important reference for greenhouse crops and because hydroponics is a technique that saves precious resources, such as water. Fuzzy control allows a person to give instructions to the greenhouse in a natural language. Therefore, we generate input variables to relate the temperature inside the greenhouse to favorable or unfavorable conditions for crop growth and an output variable that allows the control to keep the window closed or to open the window at 45 or 90 degrees. After the mathematical model was refined, it is executed in a GNU Octave environment to generate the temperature values at which the window should react.
IEEE Latin America Transactions, 2020
Scene classification is a computer vision task that aims to identify the kind of scene (such as f... more Scene classification is a computer vision task that aims to identify the kind of scene (such as forest, mountain, beach) where a picture was taken. Scene classification has application in the development of automatic surveillance systems, robotic navigation, content-based image retrieval systems among other areas. According to how a scene is recognized, scene classification algorithms can be divided in two categories: based on object recognition and based on low-level image features obtained by applying descriptors. This paper proposes a new binary descriptor called Local Lineal Binary Pattern and a new framework that allows the combination of the new binary descriptor with another local and global descriptors in order to improve the automatic classification of scene images. This new binary descriptor increases the spatial support for its calculation allowing to add more spatial information than the traditional binary descriptors such as the Local Binary Pattern and the Modified Census Transform. Experiments conducted over indoor and outdoor scene datasets show that the new proposed descriptor and framework help to improve the results obtained by related works.
International Journal of Space-Based and Situated Computing, 2017
This work presents an approach for enhancing the K-means algorithm in the classification phase. T... more This work presents an approach for enhancing the K-means algorithm in the classification phase. The approach consists in a heuristic, which at each time that an object remains in the same group, between the current and the previous iteration, it is identified as stable and it is removed from computations in the classification phase in the current and subsequent iterations. This approach helps to reduce the execution time of the standard version. It can be useful in big data applications. For evaluating computational results, both the standard and the proposal were implemented and executed using three synthetic and seven well-known real instances. After testing both versions, it was possible to validate that the proposed approach spends less time than the standard one. The best result was obtained for the transactions instance when it was grouped into 200 clusters, achieving a time reduction of 90.1% with a reduction in quality of 3.97%.
Monitoring of Cardiac Arrhythmia Patterns by Adaptive Analysis
Lecture Notes on Data Engineering and Communications Technologies, 2016
In this paper, a study and development of a monitoring adaptive system based on dynamic quadratic... more In this paper, a study and development of a monitoring adaptive system based on dynamic quadratic neural unit are presented. The system is trained with a recurrent learning method, sample-by-sample in real time. This model will help to the prediction of possible cardiac arrhythmias in patients between 23 to 89 years old, age range of the electrocardiogram signals obtained from the Massachusetts Institute of Technology-Beth Israel Hospital arrhythmia database. By means of the implementation of this adaptive monitoring system the model is capable of processing heart rate signals in real time and to recognize patterns that predict cardiac arrhythmias up to 1 second ahead. The Dynamic Quadratic Neural Unit in real time has demonstrated presenting greater efficiency and precision comparing with multilayer perceptron-type neural networks for pattern classification and prediction; in addition, this architecture has demonstrated in developed research, to be superior to other different type of adaptive architectures.
2008 7th IEEE International Conference on Cognitive Informatics, 2008
As the neural-symbolic hybrid systems (NSHS) gain acceptance, it increases the necessity to guara... more As the neural-symbolic hybrid systems (NSHS) gain acceptance, it increases the necessity to guarantee the automatic validation and verification of the knowledge contained in them. In the past, such processes were made manually. In this paper, an enhanced Petri net model is presented to the detection and elimination of structural anomalies in the knowledge base of the NSHS. In addition,
2014 10th International Conference on Natural Computation (ICNC), 2014
This paper presents a predictive model for the prediction and modeling of nonlinear, chaotic, and... more This paper presents a predictive model for the prediction and modeling of nonlinear, chaotic, and non-stationary electrocardiogram signals. The model is based on the combined usage of Hilbert-Huang transform, False nearest neighbors, and a novel neural network architecture. This model is intended to increase the prediction accuracy by applying the Empirical Mode Decomposition over a signal, and to reconstruct the signal by adding each calculated Intrinsic Mode Function and its residue. The Intrinsic Mode Function that obtains the highest frequency oscillation is not considered during the reconstruction. The optimal embedding dimension space of the reconstructed signal is obtained by False Nearest Neighbors algorithm. Finally, for the prediction horizon, a neural network retraining technique is applied to the reconstructed signal. The method has been validated using the record 103 from MIT-BIH arrhythmia database. Results are very promising since the measured root mean squared errors are 0.031, 0.05, and 0.085 of the ECG amplitude, for the prediction horizons of 0.0028, 0.0056, 0.0083 seconds, respectively.
Adaptive Threshold, Wavelet and Hilbert Transform for QRS Detection in Electrocardiogram Signals
This paper combines Hilbert and Wavelet transforms and an adaptive threshold technique to detect ... more This paper combines Hilbert and Wavelet transforms and an adaptive threshold technique to detect the QRS complex of electrocardiogram signals. The method is performed in a window framework. First, the Wavelet transform is applied to the ECG signal to remove noise. Next, the Hilbert transform is applied to detect dominant peak points in the signal. Finally, the adaptive threshold technique is applied to detect R-peaks, Q, and S points. The performance of the algorithm is evaluated against the MIT-BIH arrhythmia database, and the numerical results indicated significant detection accuracy.
Information Streaming Systems: A Review
2018 Innovations in Intelligent Systems and Applications (INISTA), 2018
Ubiquitous nature of information determines continuous decision making always based on its stream... more Ubiquitous nature of information determines continuous decision making always based on its stream coming from the environment, not only in the human, but also in the animal world. Communication between the users of information system is a fundamental concept for acquiring, showing, spreading, sharing and constantly increasing the knowledge about the circumstances under which they are in need to make a move all alone as well as while working in groups. In this paper, we review information stream-based systems on amount of information acceptable to facilitate the process of decision making, also in multilingual settings. We analyze the information system objects and the resource management cycle. In addition, methodology of data mining in natural language is investigated. Multilingual information flow between the users is specific while analyzing from the contextual perspective, when the range of information system applications becomes extended to international cooperation.
Source-target mapping model of streaming data flow for machine translation
2017 IEEE International Conference on INnovations in Intelligent SysTems and Applications (INISTA), 2017
Streaming information flow allows identification of linguistic similarities between language pair... more Streaming information flow allows identification of linguistic similarities between language pairs in real time as it relies on pattern recognition of grammar rules, semantics and pronunciation especially when analyzing so called international terms, syntax of the language family as well as tenses transitivity between the languages. Overall, it provides a backbone translation knowledge for building automatic translation system that facilitates processing any of various abstract entities which combine to specify underlying phonological, morphological, semantic and syntactic properties of linguistic forms and that act as the targets of linguistic rules and operations in a source language following professional human translator. Streaming data flow is a process of mining source data into target language transformation during which any inference impedes the system effectiveness by producing incorrect translation. We address a research problem of exploring streaming data from source-targ...
Uploads
Papers by Ricardo Rodriguez Jorge
local and global descriptors in order to improve the automatic classification of scene images. This new binary descriptor increases the spatial support for its calculation allowing to add more spatial information than the traditional binary descriptors such as the Local Binary Pattern and the Modified Census Transform. Experiments conducted over indoor and outdoor scene datasets show that the new proposed descriptor and framework help to improve the results obtained by related works.