Publications
Permanent URI for this collection
Browse
Browsing Publications by Department
Results Per Page
Sort Options
-
PublicationA biologically motivated computational architecture inspired in the human immunological system to quantify abnormal behaviors to detect presence of intrudersIn this article is presented a detection model of intruders by using an architecture based in agents that imitates the principal aspects of the Immunological System, such as detection and elimination of antigens in the human body. This model is based on the hypothesis of an intruder which is a strange element in the system, whereby can exist mechanisms able to detect their presence. We will use recognizer agents of intruders (Lymphocytes-B) for such goal and macrophage agents (Lymphocytes-T) for alerting and reacting actions. The core of the system is based in recognizing abnormal patterns of conduct by agents (Lymphocytes-B), which will recognize anomalies in the behavior of the user, through a catalogue of Metrics that will allow us quantify the conduct of the user according to measures of behaviors and then we will apply Statistic and Data Minig technics to classify the conducts of the user in intruder or normal behavior. Our experiments suggest that both methods are complementary for this purpose. This approach was very flexible and customized in the practice for the needs of any particular system. © 2006 International Federation for Information Processing.
-
PublicationA Detailed Study on the Choice of Hyperparameters for Transfer Learning in Covid-19 Image Datasets using Bayesian OptimizationFor many years, the area of health care has evolved, mainly using medical images to detect and evaluate diseases. Nowadays, the world is going through a pandemic due to COVID-19, causing a severe effect on the health system and the global economy. Researchers, both in health and in different areas, are focused on improving and providing various alternatives for rapid and more effective detection of this disease. The main objective of this study is to automatically explore as many configurations as possible to recommend a smaller starting hyperparameter space. Because the manual selection of these hyperparameters can lose configurations that generate more efficient models, for this, we present the MKCovid-19 workflow, which uses chest x-ray images of patients with COVID-19. We use knowledge transfer based on convolutional neural networks and Bayes optimization. A detailed study was conducted with different amounts of training data. This automatic selection of hyperparameters allowed us to find a robust model with an accuracy of 98% in test data.
-
PublicationA field trial of alternative targeted screening strategies for chagas disease in Arequipa, Peru( 2012-01-01)
;Hunter G. ;Bern C.Background: Chagas disease is endemic in the rural areas of southern Peru and a growing urban problem in the regional capital of Arequipa, population ~860,000. It is unclear how to implement cost-effective screening programs across a large urban and periurban environment. Methods: We compared four alternative screening strategies in 18 periurban communities, testing individuals in houses with 1) infected vectors; 2) high vector densities; 3) low vector densities; and 4) no vectors. Vector data were obtained from routine Ministry of Health insecticide application campaigns. We performed ring case detection (radius of 15 m) around seropositive individuals, and collected data on costs of implementation for each strategy. Results: Infection was detected in 21 of 923 (2.28%) participants. Cases had lived more time on average in rural places than non-cases (7.20 years versus 3.31 years, respectively). Significant risk factors on univariate logistic regression for infection were age (OR 1.02; p = 0.041), time lived in a rural location (OR 1.04; p = 0.022), and time lived in an infested area (OR 1.04; p = 0.008). No multivariate model with these variables fit the data better than a simple model including only the time lived in an area with triatomine bugs. There was no significant difference in prevalence across the screening strategies; however a self-assessment of disease risk may have biased participation, inflating prevalence among residents of houses where no infestation was detected. Testing houses with infected-vectors was least expensive. Ring case detection yielded four secondary cases in only one community, possibly due to vector-borne transmission in this community, apparently absent in the others. Conclusions: Targeted screening for urban Chagas disease is promising in areas with ongoing vector-borne transmission; however, these pockets of epidemic transmission remain difficult to detect a priori. The flexibility to adapt to the epidemiology that emerges during screening is key to an efficient case detection intervention. In heterogeneous urban environments, self-assessments of risk and simple residence history questionnaires may be useful to identify those at highest risk for Chagas disease to guide diagnostic efforts. -
PublicationA Heterogeneous Scalable-Orchestration Architecture for Home AutomationInternet of Things is represented by the large number of smart devices connected to the internet and the number of devices is constantly growing, according to Gartner this amount would reach 20.4 billion devices by 2020, there are many industries that make use of this technology, among them is the home automation, where we can find several architectural proposals that try to solve the implementation of devices connected to the Internet in an environment, however most of these proposals only give solutions for specific requirements using a certain technology, avoiding problems such as network management, security, heterogeneous devices scalability, etc. In this paper we present a novel Architecture for Home Automation based on the guidelines proposed by ISO/IEC 30141:2018 (Internet of Things), as well as use cases proposed by the OneM2M. The implementation of this architecture will guarantee us an easy integration, orchestration of IoT devices adapting to the environment context as well as managing control in communications.
-
PublicationA new approach for supervised learning based influence value reinforcement learning(Association for Computing Machinery, 2018-02-02)
;Valdivia A.The neural self-organization, is an innate feature of the mammal's brains, and is necessary for its operation. The most known neuronal models that use this characteristic are the self-organized maps (SOM) and the adaptive resonance theory (ART), but those models, did not take the neuron as a processing unit, as the biological counterpart. On the other hand, the influence value learning paradigm [1], used in multi-agent environments, proof that agents can communicate with each other [2]; and they can self-organize to assign tasks; without any interference. Motivated by this missing feature in artificial networks, and with the influence value reinforcement learning algorithm; a new approach to supervised learning was modeled using the neuron as an agent learning by reinforcement. -
PublicationA new approach for the categorization of boats for artisanal fishing in Peru(Latin American and Caribbean Consortium of Engineering Institutions, 2018-01-01)
;Inca R.C.Paredes J.V.Artisanal fishing is one of the oldest and most important activities in Peru. Currently the fishing regulations do not contemplate an appropriate categorization to the physical characteristics presented by these vessels. In the paper a new approach for the categorization of boats is proposed crafts in Peru considering the data published on the page of the Ministry of Production, which from a filtering has a total of 6056 records, giving a categorization in 4 groups (ZAPATO, CHALANA, BOTE and LANCHA). -
PublicationA new mulinane diterpenoid from the cushion shrub Azorella compacta growing in PerĂș( 2014-01-01)
;Salgado F. ;SepĂșlveda B. ;Simirgiotis M. ;Quispe L.Background: Azorella compacta is a rare yellow-green compact resinous cushion shrub growing from the high Andes of southern Peru to northwestern Argentina, and which is a producer of biologically active and unique diterpenoids. Objective: This study investigated the secondary metabolites present in a Peruvian sample of Azorella compacta and the evaluation of gastroprotective activity of the isolated compounds in a gastric- induced ulcer model in mice. Material and Methods: Six secondary metabolites (diterpenoids 1-6) present in the dichloromethane (DCM) extract of A. compacta growing in Peru were isolated by a combination of Sephadex LH-20 permeation and silica gel chromatography and their chemical structures were elucidated by spectroscopic methods (NMR) and molecular modeling. The gastroprotective activity of the new compound 1 was evaluated on the HCl/EtOH-induced gastric lesion model in mice and compared to the activity showed by the known compounds. Results: A new mulinane diterpene along with five known diterpenoids have been isolated from a Peruvian sample of A. compacta and the gastroprotective results show that compound 1 is less active than the other known mulinane diterpenoids isolated. Conclusions: A. compacta growing in Peru showed the presence of the new mulinane 1, which was poorly active in the HCl/EtOH-induced gastric lesion model in mice. Indeed, the activity was lower than other diterpenoids (2-6) showing an oxygenated function at C-16 or/and C-20, which confirm the role of an oxygenated group (OH or carboxylic acid) for the gastroprotective activity of mulinane compounds. -
PublicationA novel fuzzy probabilistic clustering algorithm for satellite image segmentation(Institute of Electrical and Electronics Engineers Inc., 2018-10-12)Satellite Image Segmentation is a task widely investigate since we can extract and analyze information of an image. In satellite image, the information of each one of the bands must be considered. We propose a new method based on the New Fuzzy Centroid Model and includes spatial information. Furthermore, we use the occurrence of each intensity value in a particular band and the Gaussian function in order to compute the degree of contribution of pixels in the neighborhood. By incorporating spatial information (global and local), we improve the clustering process and consequently, a better segmentation is obtained. This paper reports preliminary results of experiments that show that the proposed algorithm performs accurately on a real data set. For the evaluation of the algorithm, different cluster validity indexes are employed.
-
PublicationA predictive model implemented in knime based on learning analytics for timely decision making in virtual learning environments(International Journal of Information and Education Technology, 2022-02-01)
;Valderrama-Chauca E.D. ;Cari-Mogrovejo L.H. ;Apaza-Huanca J.M.Sanchez-Ilabaca J.The present research aims to implement a predictive model in the KNIME platform to analyze and compare the prediction of academic performance using data from a Learning Management System (LMS), identifying students at academic risk in order to generate timely and timely interventions. The CRISP-DM methodology was used, structured in six phases: Problem analysis, data analysis, data understanding, data preparation, modeling, evaluation and implementation. Based on the analysis of online learning behavior through 22 behavioral indicators observed in the LMS of the Faculty of Educational Sciences of the National University of San Agustin. These indicators are distributed in five dimensions: Academic Performance, Access, Homework, Social Aspects and Quizzes. The model has been implemented in the KNIME platform using the Simple Regression Tree Learner training algorithm. The total population consists of 30,000 student records from which a sample of 1,000 records has been taken by simple random sampling. The accuracy of the model for early prediction of students' academic performance is evaluated, the 22 observed behavioral indicators are compared with the means of academic performance in three courses. The prediction results of the implemented model are satisfactory where the mean absolute error compared to the mean of the first course was 3. 813 and with an accuracy of 89.7%, the mean absolute error compared to the mean of the second course was 2.809 with an accuracy of 94.2% and the mean absolute error compared to the mean of the third course was 2.779 with an accuracy of 93.8%. These results demonstrate that the proposed model can be used to predict students' future academic performance from an LMS data set. -
PublicationA revision of Neusticomys peruviensis (Rodentia: Cricetidae) with the description of a new subspecies(Oxford University Press, 2020)
;SĂĄnchez-VendizĂș P.Neusticomys peruviensis is a poorly known sigmodontine rodent of the tribe Ichthyomyini, represented in collections by only five specimens collected in five localities from lowland forests of central and southern Peru. Recent expeditions in Llanchama, in northern Peru, north of the RĂo Amazonas, and near Allpahuayo Mishana Natural Reserve (Loreto, Peru), were successful in obtaining three specimens of Neusticomys. Based on morphological and meristic data, we found the population at Llanchama is distinct from the allopatric populations of N. peruviensis, and other species of Neusticomys. A species distribution model also shows the population at Llanchama is not highly predicted by the set of variables of the known localities of N. peruviensis. However, sequence data from the mitochondrial cytochrome-b gene indicate that genetic distinctiveness is low. Because intraspecific variability is important to understand evolutionary and biogeographic processes, and in concordance with the polytypic species concept, we interpret the population at Llanchama to represent a new subspecies of N. peruviensis that we describe in this paper. -
PublicationA transformational model for the prediction of the S&P 500 stock index using fuzzy neural networks and genetic algorithms(Latin American and Caribbean Consortium of Engineering Institutions, 2017)
;Javier-Quispe D. ;Valdez-Yana G. ;Vargas-HuamĂĄn J.The objective of this paper is the elaboration of a transformational model to forecasting the diary variance of S&P 500 by using genetic algorithms and fuzzy neural networks. The model consists of two phases, the first is the elaboration of fuzzy functions and rules through TSK-IRL-R in KEEL, and the second is the training of fuzzy neural network by using ANFIS in MATLAB. The data set was composed the diary variances from Yahoo Finance. It was obtained 7.5843 of training error. -
PublicationAcademic impact of covid-19 on peruvian university students(Editorial Ciencias Medicas, 2021-01-01)
;Franco Rodriguez-Alarcon J. ;Charri J.C. ;Liendo-Venegas D. ;Morocho-Alburqueque N. ;Benites-Ibarra C.A. ;Avalos-Reyes M.S. ;Medina-Palomino D.S. ;Carranza-Esteban R.F.Introduction: As a result of the coronavirus pandemic, many students are worried that they may lose their academic term. Objective: Validate a scale to measure the perception of possible academic impacts among Peruvian university students. Method: A validation process was conducted of a test measuring the perception of university students about possible academic impacts. The validation was based on a scientific bibliography search, development of a preliminary overview, validation of the test by 59 experts: epidemiologists, researchers and physicians, exploratory factor analysis, and statistical analysis. Results: In terms of relevance, item 7 was found to be more essential or important than the others (V = 1.00; CI 95 %: 0.73-0.96). Item 8 (M = 3.80; Ï = 1.152) exhibits the highest mean, and item 6 the lowest (M = 3.03; SD = 1.342). AFE relevance is justified by the KMO index (0.85) and Bartlett's test (12577.0; gl = 28; p = < 0.001), both of which were found to be acceptable and significant. A Cronbach's α coefficient of 0.899 was obtained, with a confidence interval of 95%, SD 0.882-0.898, indicating a good consistency level. Conclusions: A single factor scale was validated which measures the perception of university students about the possible impacts of the coronavirus pandemic on their studies. It is necessary to evaluate each reality, and the tool may serve as a base scale for that purpose. -
PublicationAlgorithm and framework for tower fault caused by ice overloadIn the last years, energy transmission companies has challenges in ice conditions caused by climatic change around the world. Overhead lines could be affected for the ice overload over the cable. Especially in Andean mountains higher than 4500 m above mean sea level (mamsl). The target of this study is to establish the root cause analysis and conditions, for towers collapses located in the Andean mountains and the recommendations at the design stage. It is a contribution to new transmission systems associated to climate change; it has dramatically changed the cold and freezing stages in the Peruvian Andes, if this new considerations has not been implemented, then, the towers could be in a degradation process at the near future.
-
PublicationAn improve to human computer interaction, recovering data from databases through spoken natural languageThe fastest and most straightforward way of communication for mankind is the voice. Therefore, the best way to interact with computers should be the voice too. That is why at the moment men are searching new ways to interact with computers. This interaction is improved if the words spoken by the speaker are organized in Natural Language. In this article, it is proposed a model to recover information from databases through queries in Spanish Natural Language using the voice as the way of communication. This model incorporates a Hybrid Intelligent System based on Genetic Algorithms and a Kohonen Self-Organizing Map (SOM) to recognize the present phonemes in a word through time. This approach allows us to remake up a word with speaker independence. Furthermore, it is proposed the use of a compiler with type 2 grammar according to the Chomsky Hierarchy to support the syntactic and semantic structure in Spanish language. Our experiments suggest that the Spoken Natural Language improves notably the Human-Computer interaction when compared with traditional input methods such as: mouse or keybord. © Springer-Verlag Berlin Heidelberg 2007.
-
PublicationAnalysis of Player User Experience with Learning Objects in a Gamified Educational Platform: Mathematical Case(Association for Computing Machinery, 2021-07-23)
;Rey Cayro Mamani A.Alberto Deza Veliz D.E-learning and virtual platforms are gaining relevance, as they are an alternative for distance education and personalization. It is proposed to evaluate the use and modeling of the elements of gamification mechanics through the interaction of students with Learning Objects (LO), in solving problems in mathematics and the effects generated in the level of learning achievement. This article presents a quantitative approach and a correlational descriptive design to characterize the types of users and their behavior in a gamified educational platform, for second year high school students. The results show an inclination of the user to the type of "player", prioritized within the preferences of students between 12 and 14 years old by the motivations of obtaining virtual money through the accumulation of points. Likewise, it reflects a significant integration of the interactions of the students with the LO, consequently it is a valid process for the use and modeling with gamification mechanics in the educational context, attending the school curriculum and evidencing the learning achievement obtained and responsiveness in the interaction with the proposed LO. -
PublicationAnalysis of the academic performance of systems engineering students, desertion possibilities and proposals for retention(Universidad de Tarapaca, 2020-01-01)Problems with late studies and desertion affect educational institutions, students who have them, and their families, hence the importance of studying them. In this work, we analyze the academic performance of the 2011-2016 cohorts of the Professional School of Systems Engineering of a public university, for this we have information from 976 students: university admission score, subjectsâ qualifications and some personal data. The overall academic performance and the first year of study are statistically described. The adjusted weighted average and exogenous, endogenous, and total performance rates are calculated. With this set of variables, using a proprietary app and a commercial app, data mining techniques are applied to find patterns that describe student academic behavior. By applying classification techniques: Neural networks and decision trees, it is found that the most influential variables are the exogenous performance rate and the ratio of approved credits in relation to the credits that in theory had to be approved; for this the CRISP-DM methodology is used. As a result, some strategies are proposed that could decrease the studied problems. It is concluded that when analyzing a studentâs academic performance, the qualifications obtained are not sufficient, their academic behavior, their performance in relation to their cohort and the pace of progress in the approval of the subjects should be considered. Thus, the techniques used allow students at academic risk to be identified in a timely manner.
-
PublicationApplication of a fuzzy decision tree with ambiguity classification to determine excess weight in schoolchildren(Sociedad Mexicana de Ingenieria Biomedica, 2018)The decision tree technique in the health sciences serves to understand the correlations between the descriptions of patients and to classify accurately in various categories. The aim of the study was to analyze the accuracy of the classification of excess weight of schoolchildren through the application of a fuzzy decision tree, using a database of ItaupĂș, ParanĂĄ (Brazil). We used the database of a sample consisting of 5962 students (3024 female and 2938 male), with an age range between 6 to 17 years of age. The variables considered were weight, height and the Body Mass Index (BMI). To classify the anthropometric data of the students, a diffuse decision tree was used. The learning results showed a correct classification in the female sex of 2688 and in the male sex of 2471 records respectively. In relation to accuracy, 84% was determined in the male sex and 89% in the female sex. The Area under the curve showed higher values in the Fuzzy method and in both sexes (0.965-0.983), while in the classical method, they were lower (0.804-0.895). According to the calculated results it is possible to apply the fuzzy decision tree for the classification of overweight students with an acceptable accuracy, and it is presented as an alternative technique that can save time when analyzing the nutritional status, however, no other statistical calculations were made that have to do with the precision and accuracy through conventional statistical methods and compare with the technique of fuzzy trees.
-
PublicationApplication of Kinect technology and artificial neural networks in the control of rehabilitation therapies in people with knee injuries(Science and Information Organization, 2020)
;Loayza B.G. ;Bendita A.C. ;Cjuno M.H.In the field of physiotherapy, the recognition of the poses of the human body is obtaining more research so that the patient has an accelerated recovery rate in his rehabilitation. Nowadays, it is not so challenging to have devices like Microsoft Kinect that allow us to interact with the user for the recognition of poses and body gestures. The objective of this work to capture the data of the joints of a person's body through a set of angles using the Kinect device, then artificial neural networks with the Back-Propagation algorithm were used for machine learning, and their precision was determined. The results found on the performance of the neural network show that 99.70% accuracy was achieved in the classification of the patients' postures, which can be used as an alternative in the rehabilitation therapies of patients with knee injuries. -
PublicationApplication of the ANFIS Neuro-Fuzzy model for the classification of obesity in children and adolescents(Latin American and Caribbean Consortium of Engineering Institutions, 2018)
;Soto-Paredes C. ;CĂĄrdenas-Soria R. ;Huancco-Coila L.The objective of this article is to classify obesity in boys and adolescents, between 6 and 17 years old, using Neural Networks and Fuzzy Logic. The neuro-diffuse model ANFIS (Fuzzy Inference System of the Artificial Neural Network) was chosen, which is in the toolbox of Matlab. ANFIS includes a complete set of features for both the fuzzification, defuzzification, training and testing. Experimental tests show a 96.96% accuracy in classification and 3.04% error. -
Publication