Search

Unsatisfactory public services encourage the public to submit complaints/ reports to public service providers to improve their services. However, each complaint/ report submitted varies. Therefore, the first step of the community complaint resolution process is to classify every incoming community complaint. The Ombudsman of The Republic of Indonesia annually receives a minimum of 10,000 complaints with an average of 300-500 reports per province per year, classifies complaints/ community reports to divide them into three classes, namely simple reports, medium reports, and heavy reports. The classification process is carried out using a weight assessment of each complaint/ report using 5 (five) attributes. It becomes a big job if done manually. This impacts the inefficiency of the performance time of complaint management officers. As an alternative solution, in this study, a machine learning method with the Naïve Bayes Classifier algorithm was applied to facilitate the process of automatically classifying complaints/ community reports to be more effective and efficient. The results showed that the classification of complaints/ community reports by applying the Naïve Bayes Classifier algorithm gives a high accuracy value of 92%. In addition, the average precision, recall, and f1-score values, respectively, are 91%, 93%, and 92%.

Social media has developed rapidly today, so social media is no longer just a place to interact and socialize but also to express opinions or criticize a particular party or institution. After the incident at the Malang Kanjuruhan stadium in October 2022, many netizens criticized the performance of PSSI as Indonesia's number one organization that oversees football competitions in Indonesia. For this reason, sentiment analysis was carried out on the official PSSI account on Twitter to assess the performance of PSSI by grouping them as Satisfied and Unsatisfied using the Naïve Bayes Classifier. Sentiment analysis took tweets from the official PSSI account and as many as 1000 comments to be used as a dataset. Then preprocessing is carried out in the GATA Framework using the Annotation Removal, Remove Hashtag, Transformation Remove URL, Regexp, Indonesian Steaming, and Indonesian Stopword Removal methods. The results obtained were 82.82% for accuracy, 78.69% for precision, 90.33% for recall, and 0.866 for AUC. With these results, the value obtained is at a good classification level.

The information retrieved from the academic texts such as articles, proceedings, thesis and project proposals are used for a wide range of purposes. In the first phase of this study; a library, that can transform the raw text into a standard form, is created by considering the key terms/features in the engineering field. Then, the key terms that can best represent the document are retrieved and a similarity detection algorithm is developed using these terms. Finally, the Naïve Bayes Classifier in machine learning is used to assign the documents to the appropriate engineering sub-fields. The project proposals submitted to TUBITAK Academic Research Funding Program Directorate (ARDEB) are analyzed as a case study. The results indicate that the proposed similarity algorithm correctly detects almost all of the revised proposals while the accuracy of the classifier is 83.3% in the first prediction and reaches up to 96.4% in the first three predictions over a sample of 1255 proposals. [ABSTRACT FROM AUTHOR]

Copyright of Polish Statistician / Wiadomości Statystyczne is the property of State Treasury - Statistics Poland and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Along with the growth of the Internet comes a competitive environment among Internet service providers. In this regard, quality of service (QoS) and customer's quality of experience (QoE) are introduced as the two main criteria of satisfaction for network users/regulators. In this paper, we evaluate these two criteria for one of the largest Internet service communication networks in Iran. By providing a predictive model, we propose a solution to improve the quality of the communication. Our model predicts the quality of experience from the quality-of-service parameters with an accuracy of roughly 90%. Next, we reverse-engineer the relationship between the quality of experience and quality of service to develop a service level agreement (SLA) contract. The relationship between quality of experience and quality of service is then compiled into a set of if-then rules. By using a decision tree classifier, we were able to set the quality-of-service parameter thresholds for the gold, silver, and bronze SLAs. [ABSTRACT FROM AUTHOR]

Aim: Sepsis is a leading cause of morbidity and mortality in neonates. Early diagnosis is key but difficult due to non‐specific signs. We investigate the predictive value of machine learning‐assisted analysis of non‐invasive, high frequency monitoring data and demographic factors to detect neonatal sepsis. Methods: Single centre study, including a representative cohort of 325 infants (2866 hospitalisation days). Personalised event timelines including interventions and clinical findings were generated. Time‐domain features from heart rate, respiratory rate and oxygen saturation values were calculated and demographic factors included. Sepsis prediction was performed using Naïve Bayes algorithm in a maximum a posteriori framework up to 24 h before clinical sepsis suspicion. Results: Twenty sepsis cases were identified. Combining multiple vital signs improved algorithm performance compared to heart rate characteristics alone. This enabled a prediction of sepsis with an area under the receiver operating characteristics curve of 0.82, up to 24 h before clinical sepsis suspicion. Moreover, 10 h prior to clinical suspicion, the risk of sepsis increased 150‐fold. Conclusion: The present algorithm using non‐invasive patient data provides useful predictive value for neonatal sepsis detection. Machine learning‐assisted algorithms are promising novel methods that could help individualise patient care and reduce morbidity and mortality. [ABSTRACT FROM AUTHOR]

Intelligent connected vehicles (ICVs) constitute a transformative technology attracting immense research effort and holding great promise in providing road safety, transport efficiency, driving comfort, and eco-friendly mobility. As the driving environment becomes more and more "connected", the manner in which an ICV is driven (driving style) can dynamically vary from time to time, due to the change in several parameters associated with personal traits and with the ICV's surroundings. This necessitates fast and effective decisions to be made for a priori identifying the most appropriate driving style for an ICV. Accordingly, the main goal of this study is to present a novel, in-vehicle autonomous decision-making functionality, which enables ICVs to dynamically, transparently, and securely utilize the best available driving style (DS). The proposed functionality takes as input several parameters related to the driver's personal characteristics and preferences, as well as the changing driving environment. A Naive Bayes learning classifier is applied for the cognitive nature of the presented functionality. Three scenarios, with regards to drivers with different personal preferences and to driving scenes with changing environment situations, are illustrated, showcasing the effectiveness of the proposed functionality. [ABSTRACT FROM AUTHOR]

In this paper, we propose a new computationally fast method for text-based image retrieval from unlabeled galleries, where retrieval is formulated as a multi-class learning problem. While most existing methods assign images representing the same concept with equal importance during learning, we propose a weighted multi-view likelihood term to deal with the intra-class variations within training set of each concept. At first, we cluster each training set to detect the concept's visual appearances (views). Because number of clusters may significantly vary from one set to another, abusively unifying such a hyper-parameter over all the sets could degrade the learning outcomes. We, therefore, propose to automatically and precisely accomplish this task using Davies-Bouldin index. Noting that images are represented using deep features, which are normalized using vanilla-L2 rule to deal with bursty visual features. The proposed multi-view term is constructed by combining multivariate normal probability density functions related to the resulting clusters. This term is then incorporated within a naïve Bayes classifier alongside with the prior probability of the concept, where each component is weighted using Expectation-Maximization (EM) algorithm. Given a textual query, relevant images are the ones that reach the maximum scores of posterior probability, which is calculated using our Bayesian learning scheme. Experimental results on public datasets demonstrate the effectiveness and rapidity of the proposed method compared to several other methods. [ABSTRACT FROM AUTHOR]

We examine the significance of the predictive potential of EPI cystatin C (EPI CysC) in combination with NTproBNP, sodium, and potassium in the evaluation of renal function in patients with cardiorenal syndrome using standard mathematical classification models from the domain of artificial intelligence. The criterion for the inclusion of subjects with combined impairment of heart and kidney function in the study was the presence of newly discovered or previously diagnosed clinically manifest cardiovascular disease and acute or chronic kidney disease in different stages of evolution. In this paper, five standard classifiers from the field of machine learning were used for the analysis of the obtained data: ensemble of neural networks (MLP), ensemble of k-nearest neighbors (k-NN) and naive Bayes classifier, decision tree, and a classifier based on logistic regression. The results showed that in MLP, k-NN, and naive Bayes, EPI CysC had the highest predictive potential. Thus, our approach with utility classifiers recognizes the essence of the disorder in patients with cardiorenal syndrome and facilitates the planning of further treatment. [ABSTRACT FROM AUTHOR]

This research aims to analyze the performance of three classification models, namely Decision Tree Classifier, Support Vector Machine, and Naive Bayes Classifier, in predicting lung cancer using the "Lung Cancer Prediction" dataset. The performance evaluation metrics used include accuracy, precision weighted, recall weighted, and F1 weighted. As a preliminary step, exploratory data analysis (EDA) and dataset preprocessing, including feature selection, data cleaning, and data transformation, were conducted. The test data results showed that the Decision Tree Classifier and Naive Bayes Classifier had similar performances with high accuracy, precision, recall, and F1 values. Meanwhile, the Support Vector Machine also exhibited competitive performance, although its precision weighted value was slightly lower. Additionally, an outlier analysis was conducted using box plots, revealing that the Decision Tree Classifier had 2 outlier values, while the Support Vector Machine had 4 outlier values, and Naive Bayes had no outlier values. In conclusion, all three classification models demonstrated good potential in lung cancer prediction. However, selecting the best model requires consideration of relevant evaluation metrics for the application and accommodating the limitations of each model. Further evaluation and in-depth analysis are needed to ensure the reliability of the models in predicting lung cancer cases more accurately and consistently.

Introduction: Landslide is one of the most widespread geohazards around the world. Therefore, it is necessary and meaningful to map regional landslide susceptibility for landslide mitigation. In this research, landslide susceptibility maps were produced by four models, namely, certainty factors (CF), naive Bayes (NB), J48 decision tree (J48), and multilayer perceptron (MLP) models.Methods: In the first step, 328 landslides were identified via historical data, interpretation of remote sensing images, and field investigation, and they were divided into two subsets that were assigned different uses: 70% subset for training and 30% subset for validating. Then, twelve conditioning factors were employed, namely, altitude, slope angle, slope aspect, plan curvature, profile curvature, TWI, NDVI, distance to rivers, distance to roads, land use, soil, and lithology. Later, the importance of each conditioning factor was analyzed by average merit (AM) values, and the relationship between landslide occurrence and various factors was evaluated using the certainty factor (CF) approach. In the next step, the landslide susceptibility maps were produced based on four models, and the effect of the four models were quantitatively compared by receiver operating characteristic (ROC) curves, area under curve (AUC) values, and non-parametric tests.Results: The results demonstrated that all the four models can reasonably assess landslide susceptibility. Of these four models, the CF model has the best predictive performance for the training (AUC=0.901) and validating data (AUC=0.892).Discussion: The proposed approach is an innovative method that may also help other scientists to develop landslide susceptibility maps in other areas and that could be used for geo-environmental problems besides natural hazard assessments.

An increasing number of methods are being used to extract rock discontinuities from 3D point cloud data of rock surfaces. In this paper, a new method for automatic extraction of rock discontinuity based on an improved Naive Bayes classifier is proposed. The method first uses principal component analysis to find the normal vectors of the points, and then generates a certain number of random point sets around the selected training points for training the classifier. The trained, improved Naive Bayes classifier is based on point normal vectors and is able to automatically remove noise points due to various reasons in conjunction with the knee point algorithm, realizing high-precision extraction of the discontinuity sets. Subsequently, the individual discontinuities are segmented using a hierarchical density-based spatial clustering method with noise application. Finally, the PCA algorithm is used to complete the orientation by plane fitting the individual discontinuities. The method was applied in two cases, Kingston and Colorado, and the reliability and advantages of the new method were verified by comparing the results with those of previous research, and the discussion and analysis determined the optimal values of the relevant parameters in the algorithm.

People may now express their thoughts and ideas with a wider audience because of the popularity of social media sites like Twitter, Instagram, and Facebook. Businesses now utilise Twitter to reply to client comments, reviews, and grievances. Every day, millions of individuals discuss a wide range of issues on Twitter by sharing their ideas and interests. Sentiment analysis is a useful method for analysing such data, which involves identifying the sentiment of the source text and classifying it as positive, neutral, or negative. However, due to the vast amount of data, it can be challenging for businesses to address every customer’s question or complaint in a timely manner. Some issues may be urgent but delayed due to the volume of information. In order to prioritize emergency tweets, a system is proposed that utilizes machine learning algorithms such as Random Forest, Support Vector Machine, Logistic Regression, and Naïve Bayes to identify tweets based on their urgency. The proposed system gathers and preprocesses unstructured data, performs feature extraction, trains, assesses and compares multiple machine learning models to determine the best classifier with the highest accuracy, and uses vectorization via a pipeline to determine the sentiment of a new tweet provided as input.

Heart failure (ARF) is a health problem that has relatively high mortality and morbidity rates in developed or developing countries, including Indonesia. In 2016, WHO stated that 17.5 million people died from cardiovascular disease, while in 2008, HF disease represented 31% of patient deaths worldwide. One of the new breakthroughs for early diagnosis is utilizing data mining techniques. In this study, the Correlated Naive Bayes Classifier (C-NBC) and Naive Bayes Classifier (NBC) algorithms are used to obtaining the best accuracy results so that they can be used for the Heart Failure dataset. Based on the results of the tests that have been carried out, it shows that the Correlated Naive Bayes Classifier (C-NBC) algorithm accuracy of 80.6% obtains higher accuracy than the Naive Bayes Classifier (NBC) algorithm of 67.5%. With the results of this study, the use of the Correlated Naive Bayes Classifier (C-NBC) algorithm can be used to diagnose patients with heart failure (heart failure) because it has a high level of accuracy and is categorized as Good Classification.

Abstract Motivation Application of chemical named entity recognition (CNER) algorithms allows retrieval of information from texts about chemical compound identifiers and creates associations with physical–chemical properties and biological activities. Scientific texts represent low-formalized sources of information. Most methods aimed at CNER are based on machine learning approaches, including conditional random fields and deep neural networks. In general, most machine learning approaches require either vector or sparse word representation of texts. Chemical named entities (CNEs) constitute only a small fraction of the whole text, and the datasets used for training are highly imbalanced. Methods and results We propose a new method for extracting CNEs from texts based on the naïve Bayes classifier combined with specially developed filters. In contrast to the earlier developed CNER methods, our approach uses the representation of the data as a set of fragments of text (FoTs) with the subsequent preparati`on of a set of multi-n-grams (sequences from one to n symbols) for each FoT. Our approach may provide the recognition of novel CNEs. For CHEMDNER corpus, the values of the sensitivity (recall) was 0.95, precision was 0.74, specificity was 0.88, and balanced accuracy was 0.92 based on five-fold cross validation. We applied the developed algorithm to the extracted CNEs of potential Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) main protease (Mpro) inhibitors. A set of CNEs corresponding to the chemical substances evaluated in the biochemical assays used for the discovery of Mpro inhibitors was retrieved. Manual analysis of the appropriate texts showed that CNEs of potential SARS-CoV-2 Mpro inhibitors were successfully identified by our method. Conclusion The obtained results show that the proposed method can be used for filtering out words that are not related to CNEs; therefore, it can be successfully applied to the extraction of CNEs for the purposes of cheminformatics and medicinal chemistry.

The logistics industry is growing very rapidly. One of big industry in Indonesia is PT. Tiki Line Nugraha Ekakurir (JNE), which has been established for 29 years. This company has an extensive network in all cities in Indonesia, with service points of 1,500 locations. JNE has an application called my JNE on Google Play, which received more than 86,000 reviews and since December 2019 only got a rating of 2.4 stars out of a total rating of 5 stars. This study is obtained to analysis JNE user review data from Google Play. The reviews used in this study totaled 1,876 classified into positive and negative sentiment classes using the Naïve Bayes Classifier algorithm and word associations were also implemented. Classification with naïve bayes classifier with 90% training data and 10% test data had the best accuracy of 85.87%. Furthermore, for the text association, information is obtained that JNE users are talking about "send", "package", "courier", "good", "application", "fast", "service", "receive", "help", and "star". Whereas in the class of negative sentiment users often talk about "send", "package", "courier", "disappointed", "service", "service", "bad", "application", "severe", and "slow".

Despite the animal models’ complexity, researchers tend to reduce the number of animals in experiments for expenses and ethical concerns. This tendency makes the risk of false-positive results, as statistical significance, the primary criterion to validate findings, often fails if testing small samples. This study aims to highlight such risks using an example from experimental regenerative therapy and propose a machine-learning solution to validate treatment effects. The example analysed was the pharmacological treatment of ear pinna punch wound healing in mice. Wound closure data analysed included eight groups treated with an epigenetic inhibitor, zebularine, and eight control groups receiving vehicle alone, of six mice each. We confirmed the zebularine healing effect for all 64 pairwise comparisons between treatment and control groups but also determined minor yet statistically significant differences between control groups in five of 28 possible comparisons. The occurrences of significant differences between the control groups, regardless of standardised experimental conditions, indicate a risk of statistically significant effects in the case a compound lacking the desired biological activity is tested. Since the criterion of statistical significance itself can be confusing, we demonstrate a machine-learning algorithm trained on datasets representing treatment and control experiments as a helpful tool for validating treatment outcomes. We tested two machine-learning approaches, Naïve Bayes and Support Vector Machine classifiers. In contrast to the Mann-Whitney U-test, indicating enhanced healing effects for some control groups receiving saline alone, both machine-learning algorithms faultlessly assigned all animal groups receiving saline to the controls.

Spread-F (SF) is one of the most important types of the ionospheric irregularities as it causes ionospheric scintillation which can severely affect the performance and reliability of communication, navigation, and radar systems. The ionosonde provides the most effective and economical way to study the ionosphere and SF. However, the manual identification of SF from an ionogram is boring and hard work. To automatically identify SF on the ionogram and extend the study of SF into the middle and low latitudes of East Asia, this paper presents a statistical analysis of SF in this region, based on the naïve Bayesian classifier. The results showed that the accuracy of automatic identification reached up to 97% on both the validation datasets and test datasets composed of Mohe, I-Cheon, Jeju, Wuhan, and Sanya ionograms, suggesting that it is a promising way to automatically identify SF on ionograms. Based on the classification results, the statistical analysis shows that SF has a complicated morphology in the middle and low latitudes of East Asia. Specifically, there is a peak of occurrence of SF in the summer in I-Cheon, Jeju, Sanya, and Wuhan; however, the Mohe station has the highest occurrence rate of SF in December. The different seasonal variations of SF might be due to the different geographic local conditions, such as the inland-coastal differences and formation mechanism differences at these latitudes. Moreover, SF occurs more easily in the post-midnight hours when compared with the pre-midnight period in these stations, which is consistent with the previous results. Furthermore, this paper extracts the frequency SF (FSF) index and range SF (RSF) index to characterize the features of SF. The results shows that the most intense FSF/RSF appeared in the height range of 220–300 km/1–7 MHz in these stations, although there are different magnitude extensions on different season in these regions. In particular, strong spread-F (SSF) reached its maximum at the equinox at Sanya, confirming the frequent SSF occurrence at the equinox at the equator and low latitudes. These results would be helpful for understanding the characteristics of SF in East Asia. [ABSTRACT FROM AUTHOR]

The Cox proportional hazard model may predict whether an individual belonging to a given group would likely register an event of interest at a given time. However, the Cox model is limited by relatively strict statistical assumptions. In this study, we propose decomposing the time-to-event variable into "time" and "event" components and using the latter as a target variable for various machine-learning classification algorithms, which are almost assumption-free, unlike the Cox model. While the time component is continuous and is used as one of the covariates, i.e., input variables for various classification algorithms such as logistic regression, naïve Bayes classifiers, decision trees, random forests, and artificial neural networks, the event component is binary and thus may be modeled using these classification algorithms. Moreover, we apply the proposed method to predict a decrease or non-decrease of IgG and IgM blood antibodies against COVID-19 (SARS-CoV-2), respectively, below a laboratory cut-off, for a given individual at a given time point. Using train-test splitting of the COVID-19 dataset ( n = 663 individuals), models for the mentioned algorithms, including the Cox proportional hazard model, are learned and built on the train subsets while tested on the test ones. To increase robustness of the model performance evaluation, models' predictive accuracies are estimated using 10-fold cross-validation on the split dataset. Even though the time-to-event variable decomposition might ignore the effect of individual data censoring, many algorithms show similar or even higher predictive accuracy compared to the traditional Cox proportional hazard model. In COVID-19 IgG decrease prediction, multivariate logistic regression (of accuracy 0.811 ), support vector machines (of accuracy 0.845 ), random forests (of accuracy 0.836 ), artificial neural networks (of accuracy 0.806 ) outperform the Cox proportional hazard model (of accuracy 0.796 ), while in COVID-19 IgM antibody decrease prediction, neither Cox regression nor other algorithms perform well (best accuracy is 0.627 for Cox regression). An accurate prediction of mainly COVID-19 IgG antibody decrease can help the healthcare system manage, with no need for extensive blood testing, to identify individuals, for instance, who could postpone boosting vaccination if new COVID-19 variant incomes or should be flagged as high risk due to low COVID-19 antibodies. [ABSTRACT FROM AUTHOR]

Improving product quality is still a challenge; therefore, this article aims to propose an approach to predict customer satisfaction. We implemented the following techniques: the SMART(-ER) method, brainstorming (BM), a Likert-scale survey, the Pareto rule, the WSM method, and the Naive Bayes Classifier. Customer expectations were obtained as part of the survey research. Based on these, we determined customers' satisfaction with the current quality of the criteria and the weights of these criteria. We then applied the Pareto rule, the WSM method, and the Naive Bayes Classifier. In the proposed approach, it was predicted that current product quality is not very satisfactory to customers; that conditioned the need for improvement actions. The originality of the study is the ability to predict customer satisfaction while taking into account the weights of this criterion. The proposed approach can be used for any product. [ABSTRACT FROM AUTHOR]

Background: In mammals, normal fertilization depends on the structural and functional integrity of the zona pellucida (ZP), which is an extracellular matrix surrounding oocytes. Mutations in ZP may affect oogenesis, fertilization and early embryonic development, which may cause female infertility. Methods: A PubMed literature search using the keywords 'zona pellucida', 'mutation' and 'variant' limited to humans was performed, with the last research on June 30, 2022. The mutation types, clinical phenotypes and pregnancy outcomes were summarized and analyzed. The naive Bayes classifier was used to predict clinical pregnancy outcomes for patients with ZP mutations. Results: A total of 29 publications were included in the final analysis. Sixty-nine mutations of the ZP genes were reported in 87 patients with different clinical phenotypes, including empty follicle syndrome (EFS), ZP-free oocytes (ZFO), ZP-thin oocytes (ZTO), degenerated and immature oocytes. The phenotypes of patients were influenced by the types and location of the mutations. The most common effects of ZP mutations are protein truncation and dysfunction. Three patients with ZP1 mutations, two with ZP2 mutations, and three with ZP4 mutations had successful pregnancies through Intracytoplasmic sperm injection (ICSI) from ZFO or ZTO. A prediction model of pregnancy outcome in patients with ZP mutation was constructed to assess the chance of pregnancy with the area under the curve (AUC) of 0.898. The normalized confusion matrix showed the true positive rate was 1.00 and the true negative rate was 0.38. Conclusion: Phenotypes in patients with ZP mutations might be associated with mutation sites or the degree of protein dysfunction. Successful pregnancy outcomes could be achieved in some patients with identified ZP mutations. Clinical pregnancy prediction model based on ZP mutations and clinical characteristics will be helpful to precisely evaluate pregnancy chance and provide references and guidance for the clinical treatment of relevant patients. [ABSTRACT FROM AUTHOR]

The existence of KAI Access from PT. KAI prove their sincerity in serving consumers in this modern era. However, many negative reviews found in Google Play Store. There has been research on the review, but the analysis stage still at document level so the aspect related to the application is not known clearly and structured. So it is necessary to do an aspect-based sentiment analysis to extract the aspects and the sentiment. This study aims to do an aspect-based sentiment analysis on user reviews of KAI Access using Naive Bayes Classifier (NBC) and Support Vector Machine (SVM), with 3 scenarios. Scenario 1 uses NBC with Multinomial Naive Bayes, scenario 2 uses SVM with default Sklearn library parameter, and scenario 3, uses SVM with hyperparameter tunning, while the data scrapped from Google Play Store. The results show the majority of user sentiment is negative for each aspect, with most discussed errors aspect shows the high system errors. The test results gives the best model from scenario 3 with an average accuracy 91.63%, f1-score 75.55%, precision 77.60%, and recall 74.47%.

The current contribution aimed to evaluate the capacity of the naive Bayes classifier to predict the progression of dengue fever to severe infection in children based on a defined set of clinical conditions and laboratory parameters. This case-control study was conducted by reviewing patient files in two public hospitals in an endemic area in Mexico. All 99 qualifying files showed a confirmed diagnosis of dengue. The 32 cases consisted of patients who entered the intensive care unit, while the 67 control patients did not require intensive care. The naive Bayes classifier could identify factors predictive of severe dengue, evidenced by 78% sensitivity, 91% specificity, a positive predictive value of 8.7, a negative predictive value of 0.24, and a global yield of 0.69. The factors that exhibited the greatest predictive capacity in the model were seven clinical conditions (tachycardia, respiratory failure, cold hands and feet, capillary leak leading to the escape of blood plasma, dyspnea, and alterations in consciousness) and three laboratory parameters (hypoalbuminemia, hypoproteinemia, and leukocytosis). Thus, the present model showed a predictive and adaptive capacity in a small pediatric population. It also identified attributes (i.e., hypoalbuminemia and hypoproteinemia) that may strengthen the WHO criteria for predicting progression to severe dengue.