publications | Edwin Puertas

2024

Enhancing Vocational Guidance with Machine Learning: Predicting STEM Career Viability for High School Students

Melissa Moreno-Novoa, Edwin Puertas, and Juan Carlos Martinez-Santos

In Colombian Conference on Computing, 2024
Implementation of Convolutional Neural Networks for Automated Disease Detection in Cucumber Crops

Andrea Menco Tovar, Edwin Puertas, and Juan Carlos Martinez-Santos

In Colombian Conference on Computing, 2024
Unveiling Tourist Profiles in the Department of Sucre: A Text Analysis Approach

Danileth Almanza-Gonzalez, Edwin Puertas, and Juan Carlos Martinez-Santos

In Colombian Conference on Computing, 2024
Feature Selection for Forecasting of Energy Spot Price in the Colombian Market

Mauro A Gonzalez-Sierra, Rafael Arnedo, Edwin Puertas, and 1 more author

In 2024 IEEE ANDESCON, 2024
Automated Detection of Depression and Anxiety Using Lexical and Phonestheme Features in Spanish Texts

Elizabeth Martinez, Juan Cuadrado, Juan Carlos Martinez-Santos, and 1 more author

In IberLEF (Working Notes). CEUR Workshop Proceedings, 2024
VerbaNex AI at CLEF EXIST 2024: Detection of Online Sexism using Transformer Models and Profiling Techniques

Elizabeth Martinez, Juan Cuadrado, Juan Carlos Martinez Santos, and 1 more author

environments, 2024
VerbaNex AI at DIPROMATS 2024: Enhancing Propaganda Detection in Diplomatic Tweets with Fine-Tuned BERT and Integrated NLP Techniques

Jose Cuadrado, Elizabeth Martinez, Juan Cuadrado, and 2 more authors

2024
Verbanexai lab at homo-mex 2024: Multiclass and multilabel detection of lgbtq+ phobic content using transformers

R Gonzalez-Henao, D Marrugo-Tobon, J Martinez-Santos, and 1 more author

In Proceedings of the Iberian Languages Evaluation Forum (IberLEF 2024), co-located with the 40th Conference of the Spanish Society for Natural Language Processing (SEPLN 2024), CEUR-WS. org, 2024
Team qtb on feature selection via quantum annealing and hybrid models

Esteban Payares, Edwin Puertas, and Juan Carlos Martinez Santos

Working Notes of CLEF, 2024
A computer vision system for detecting motorcycle violations in pedestrian zones

Nicolás Hernández-Dı́az, Yersica C Peñaloza, Y Yuliana Rios, and 2 more authors

Multimedia Tools and Applications, 2024
VerbaNexAI Lab at SemEval-2024 Task 3: Deciphering emotional causality in conversations using multimodal analysis approach

Victor Pacheco, Elizabeth Martinez, Juan Cuadrado, and 2 more authors

In Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024), 2024
VerbaNexAI at MEDIQA-CORR: Efficacy of GRU with BioWordVec and ClinicalBERT in Error Correction in Clinical Notes

Juan Pajaro, Edwin Puertas, David Villate, and 2 more authors

In Proceedings of the 6th Clinical Natural Language Processing Workshop, 2024
VerbaNexAI Lab at SemEval-2024 Task 10: Emotion recognition and reasoning in mixed-coded conversations based on an NRC VAD approach

Santiago Garcia, Elizabeth Martinez, Juan Cuadrado, and 2 more authors

In Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024), 2024
VerbaNexAI Lab at SemEval-2024 Task 1: A Multilayer Artificial Intelligence Model for Semantic Relationship Detection

Anderson Morillo, Daniel Peña, Juan Carlos Martinez Santos, and 1 more author

In Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024), 2024

2023

Long-Term Effects of Degradation on Photovoltaic System Return on Investment

J. Cuadrado, E. Martinez, E. Puertas, and 1 more author

2023

Abs

The adoption of photovoltaic (PV) systems has increased significantly in recent years, driven by the demand for off-grid and on-grid residential and commercial applications. However, the high initial investment required for PV installations has limited their widespread adoption. Governments and marketing enterprises have implemented different strategies to promote PV systems to overcome this barrier, focusing on the return on investment (ROI) concept. However, the conventional approach uses limited economic factors to calculate the ROI. It fails to consider the impact of external factors, such as system degradation, which can vary between systems. To address this issue, we propose a new methodology to estimate the ROI of a photovoltaic system with greater accuracy. Our approach incorporates system-predicted degradation, calculated using historical meteorological data and prediction techniques. We applied this methodology to a photovoltaic system installed at the Universidad Tecnologica de Bolivar (UTB) in Cartagena and evaluated it against five different approaches. The results show that our proposed method offers a more accurate and reliable estimation of the ROI of a photovoltaic system, considering a broader range of factors. Overall, our work contributes to advancing the understanding of photovoltaic system ROI calculation and promotes using sustainable energy sources. By providing a more precise estimation of the ROI of a photovoltaic system, our methodology can help potential investors make more informed decisions and promote the adoption of clean energy sources. © 2003-2012 IEEE.
Dataset for detecting motorcyclists in pedestrian areas

N.H. Díaz, Y.C. Peñaloza, Y.Y. Rios, and 2 more authors

2023

Abs

This paper presents a semi-automated, scalable, and homologous methodology towards IoT implemented in Python for extracting and integrating images in pedestrian and motorcyclist areas on the road for constructing a multiclass object classifier. It consists of two stages. The first stage deals with creating a non-debugged data set by acquiring images related to the semantic context previously mentioned, using an embedded device connected 24/7 via Wi-Fi to a free and public CCTV service in Medellin, Colombia. Through artificial vision techniques, and automatically performs a comparative chronological analysis to download the images observed by 80 cameras that report data asynchronously. The second stage proposes two algorithms focused on debugging the previously obtained data set. The first one facilitates the user in labeling the data set not debugged through Regions of Interest (ROI) and hotkeys. It decomposes the information in the nth image of the data set in the same dictionary to store it in a binary Pickle file. The second one is nothing more than an observer of the classification performed by the user through the first algorithm to allow the user to verify if the information contained in the Pickle file built is correct. © 2023 The Author(s)
RealCheck: A Web Application for Fake News Detection Using Natural Language Processing

E. Puertas, J. Vasquez, and J.C. Martinez-Santos

2023

Abs

The spread of misinformation and disinformation online represents a significant challenge in today’s society. This paper proposes implementing a news verifier web application that uses artificial intelligence techniques to determine the veracity of news provided by users. The system consists of three stages: user input validation, selection of relevant sources, and information analysis. We trained a logistic regression model with an in-house dataset of 2000 sentences for input validation. We utilized cosine similarity and FastText for source selection. Finally, a large language model analyzes the context and coherence of selected articles to determine veracity. The transparent presentation of results fosters media literacy. While some implementation details may differ due to time constraints, this paper provides the overall methodology and architecture of the system. © 2023 IEEE.
Researcher Profile: An Automated Solution for Searching and Gathering People’s Profiles

J.C. Martinez-Santos, J. Vasquez, and E. Puertas

2023

Abs

This project presents an innovative solution to facilitate the search and analysis of professional profiles in academic platforms by automating manual processes and using advanced technologies such as Natural Language Processing (NLP) and third-party application programming interface (API). We designed an architecture based on a React/NextJS front-end and back-end with REST APIs to enable robust and scalable performance for the proposed system. Our proposed architecture and implementation plan provided a solid foundation for the development of the solution. Our initial prototype and advanced functionalities, such as sentiment analysis, keyword extraction, and personalized profile recommendations, are expected to improve the user experience and provide added value to the academic and professional community. © 2023 IEEE.
Detection of Online Sexism Using Lexical Features and Transformer

E. Martinez, J. Cuadrado, J.C. Martinez-Santos, and 1 more author

2023

Abs

Social networks are now integral to modern life, offering instant global communication. Unfortunately, this ease of connectivity has also led to the misuse of freedom, with many social media users expressing inappropriate and often sexist comments. In response, the field of natural language processing has actively sought solutions to detect and counteract such content. Our study builds upon our previous work presented during the SemEval 2023 competition. Initial results, achieved using only lexical features, prompted a thorough reevaluation. Recognizing the potential of advanced techniques, we embarked on a comprehensive exploration to enhance our approach. Leveraging our experience, we focused on a feature union strategy. By seamlessly combining the ’twitter-roberta-base-sentiment-latest’ transformer model with established lexical features, we developed a refined methodology that transcends the limitations of individual elements. This approach improved performance and demonstrated the synergistic potential of merging these distinct features. Our improved method revealed significant enhancements, underscoring the effectiveness of our methodology in boosting detection capabilities. It led to a notable improvement of approximately 20% in binary classification accuracy, a substantial leap from our previous results. The strategic blend of linguistic and transformer-based features emerged as the driving force behind this significant advancement. © 2023 IEEE.
Quantum N-Gram Language Models for Tweet Classification

E. Payares, E. Puertas, and J.C. Martinez-Santos

2023

Abs

This paper presents a novel approach to text classification using n-grams and quantum machine learning models. We demonstrate the effectiveness of our approach by applying it to classify tweets, a challenging and noisy text classification task. Specifically, we show that using quantum models can significantly improve classification performance compared to classical ones. Additionally, we leverage quantum hardware to compute higher-order n-grams (more than bi-grams) and observe a boost in computation performance. Our experimental results demonstrate the potential of quantum computing for text classification tasks and pave the way for further exploration of quantum machine learning methods in natural language processing. © 2023 IEEE.
UTB-NLP at SemEval-2023 Task 3: Weirdness, Lexical Features for Detecting Categorical Framings, and Persuasion in Online News

J. Cuadrado, E. Martinez, A. Morillo, and 4 more authors

2023

Abs

Nowadays, persuasive messages are increasingly frequent in social networks, which generates particular concern in several communities, given that persuasion seeks to guide others towards adopting ideas, attitudes, or actions that they consider to be beneficial to themselves. The efficient detection of news genre categories, detection of framing, and detection of persuasion techniques require several scientific disciplines, such as computational linguistics and sociology. Here we illustrate how we use lexical features given a news article to determine whether it is an opinion piece, aims to report factual news, or is satire. This paper presents a novel strategy for communication based on Lexical Weirdness. The results are part of our participation in Sub-Tasks 1 and 2 in SemEval 2023 Task 3. © 2023 Association for Computational Linguistics.
Team UTB-NLP at FinancES 2023: Financial Targeted Sentiment Analysis Using a Phonestheme Semantic Approach

J. Cuadrado, E. Martinez, J.C. Martinez-Santos, and 1 more author

2023

Abs

Sentiment analysis in the financial domain is a challenging task that plays a crucial role in understanding public opinion, monitoring market trends, and assessing the impact of news on economic agents. In this shared task, we address targeted sentiment analysis in the financial domain, focusing on identifying the main economic target in news headlines and determining the sentiment polarity towards such targets. We propose a methodology that combines transformer-based models and phonestheme embeddings to extract meaningful features from the text, which are then used in a support vector machine (SVM) classifier for sentiment classification. Our approach shows promising results, outperforming the baseline with an F1-score of 0.529229 in Task 1. This research contributes to financial sentiment analysis by addressing the complexity of financial language and considering multiple economic agents’ perspectives. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Natural Language Content Evaluation System For Multiclass Detection of Hate Speech in Tweets Using Transformers

D.A. Marrugo-Tobón, J.C. Martinez-Santos, and E. Puertas

2023

Abs

In natural language processing, accurate categorization of tweets, including detecting hate speech, plays a pivotal role in efficient information organization and analysis. This paper presents a Natural Language Contents Evaluation System specifically tailored for multi-class tweet categorization, focusing on hate speech detection. Our system enhances classification accuracy and efficiency by harnessing the power of Transformers, namely BERT and DistilBERT. By leveraging feature extraction techniques, we capture pertinent information from tweets, enabling practical analysis, categorization, and identification of hate speech instances. During training, we also tackle imbalanced corpora by employing techniques to ensure fair representation of different tweet categories, including hate speech. Our system achieves impressive accuracy through extensive training of 95%, showcasing Transformers’ effectiveness in comprehending and categorizing tweets, including identifying hate speech. Furthermore, our system maintains a good accuracy during testing of 83%, highlighting the robustness and generalizability of the trained models for hate speech detection. This system contributes to advancing automated tweet categorization, specifically in hate speech detection, providing a reliable and efficient solution for organizing and analyzing diverse tweet datasets. © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Intelligent system to detect violations in pedestrian areas committed by vehicles in the City of Cartagena de Indias

N. Hernández-Díaz, Y.C. Pañaloza, Y.Y. Rios, and 2 more authors

2023

Abs

This article presents applied research which proposes an approach to a low-cost electronic system that allows the detection of violations in pedestrian areas. For this purpose, we proposed an electronic system that implements artificial intelligence algorithms and Internet of Things architectures to detect in real-time these events. The proposal focuses on experimental laboratory work through which we developed and validated a system by analyzing a real pedestrian zone scenario and invasion of the pedestrian zone. As a result of this work, we deliver a labeled dataset to the scientific community that can serve as input to new projects and the development of better offender detection algorithms. © 2023 Latin American and Caribbean Consortium of Engineering Institutions. All rights reserved.
Component Analysis to Design an Integral Sustainability Model for Micro, Small, and Medium-Sized Enterprises

J. Vásquez, and E. Puertas

2023

Abs

Micro, small, and medium enterprises (MSMEs) are the engine of economic, environmental, and social stability in many countries but recently MSMEs have been facing instability in their sales and, production, and feature a decrease in employment that generates negative growth in the gross domestic product in emerging countries. Also, along with governmental entities, they are expected to be aware of the importance of developing sustainable models to improve production systems within the framework of sustainable development. Considering there is abundant literature on multiple sustainable models, their effective implementation in production systems is a debate topic as there is no standard guide to the processes and activities that should be followed in their adoption. The objective of this study is to design an integral sustainability model for MSMEs through the analysis of the interrelationships and dependencies of sustainability factors and a perception analysis evaluating 327 MSMEs in Colombia, using the design science research and case study as a research method. The analysis of the results has been exposed as results of a doctoral thesis which indicates that the MSMEs that were studied presented significant variations and similarities in the use of the adoption of sustainable strategies, which allowed the identification of weaknesses and strengths to focus efforts on the development of productive systems aligned with environmental sustainability. © 2023 Latin American and Caribbean Consortium of Engineering Institutions. All rights reserved.

2022

Cost-Effectiveness Analysis of Strategies of COVID-19 Vaccination in Colombia: Comparison of High-Risk Prioritization and No Prioritization Strategies With the Absence of a Vaccination Plan

G. Morales-Zamora, O. Espinosa, E. Puertas, and 7 more authors

2022

Abs

Objectives: Our study compares two national COVID-19 vaccination plan strategies—high-risk prioritization and no prioritization—and estimates their cost-effectiveness compared with no vaccination, to generate possible recommendations for future vaccination plans. Methods: We developed a Markov discrete-time, compartmental, deterministic model stratified by Colombian departments, healthcare workers, comorbidities, and age groups and calibrated to seroprevalence, cases, and deaths. The model simulates three scenarios: no vaccination, no prioritization of vaccination, and prioritization of high-risk population. The study presents the perspective of the health system of Colombia, including the direct health costs financed by the government and the direct health outcomes related to the infection. We measured symptomatic cases, deaths, and costs for each of the three scenarios from the start of the vaccination rollout to February 20, 2023. Results: Both for the base-case and across multiple sensitivity analyses, the high-risk prioritization proves to be the most cost-effective of the considered strategies. An increment of US255 million results in an incremental cost-effectiveness ratio of US3339 per disability-adjusted life-year avoided. The simulations show that prioritization of high-risk population reduces symptomatic cases by 3.4% and deaths by 20.1% compared with no vaccination. The no-prioritization strategy is still cost-effective, with an incremental cost-effectiveness ratio of US$5223.66, but the sensitivity analysis the show potential risks of losing cost-effectiveness under the cost-effectiveness threshold (one gross domestic product per averted disability-adjusted life-year). Conclusions: The high-risk prioritization strategy is consistently more cost-effective than the no-prioritization strategy across multiple scenarios. High-risk prioritization is the recommended strategy in low-resource settings to reduce the burden of disease. © 2022 International Society for Health Economics and Outcomes Research
Detection of broken bars in three-phase motors by using curve fits and classification algorithms

G. Hoyos, E. Puertas, J.L. Villa, and 1 more author

2022

Abs

Since they transform electrical energy into mechanical energy, three-phase induction motors are one of the main assets that companies have. Therefore, good monitoring of their conditions and diagnosing their faults is essential. In this article, we propose a curve fitting technique and classification algorithms for a current motor phase to detect broken bars inside the motor. The data set is in the IEEE database, where the data was acquired, simulating the conditions of healthy and broken bars by varying the load condition. The curve fitting technique gives me essential attributes such as the signal’s amplitude, frequency, and phase shift, supported by the Fourier transform, which informs how the signal power is a function of frequency. Furthermore, we extracted attributes to train the classifiers, achieving 85% accuracy in classifying the number of broken bars within the engine. © 2022 IEEE.
Presidential preferences in Colombia through Sentiment Analysis

E. Puertas, J.C. Martinez-Santos, and P. Andres Pertuz-Duran

2022

Abs

This work carries out the sentiment analysis of the social network Twitter regarding the presidential debate on May 23, where a hashtag was left open so viewers could give their points of view on these three candidates: Gustavo Petro, Federico Gutierrez, and Rodolfo Hernández. Once we extracted these Tweets contained in the hashtag, they were manually classified. They then went through all the pre-processing and elimination of special characters, links, URLs, images, or videos. Next, the TextVectorization layer from the TensorFlow library was used to convert these tweets to vectors and finally to go through the two models. The results show the best results for the BERT model with an accuracy of 76% and an F1 score of 85%. © 2022 IEEE.

2021

A sustainability maturity model for micro, small and medium-sized enterprises (MSMEs) based on a data analytics evaluation approach

J. Vásquez, S. Aguirre, E. Puertas, and 3 more authors

2021

Abs

A maturity model for micro, small and medium enterprises (MSMEs) is introduced to assess the level of implementation of sustainability strategies and practices in this type of business. According to the literature, only a few sustainability-maturity assessment models intended for MSMEs have integrated the following three factors: environmental knowledge management, environmental strategies and good practices, and environmental management systems. The sustainability maturity model proposed here for MSMEs is capable of supporting the efforts of companies in their attempts to achieve both environmental sustainability and an improvement in their production systems. The model encompasses a four-level qualitative scale and uses supervised classification algorithms to categorize companies through data analysis techniques. After applying the model to a group of MSMEs from different productive sectors in Colombia, the results have shown that 6% of the companies were at an insufficient level, 31% were at an initial level, 45% at a developed sustainability maturity level, and 18% at a consolidated level. This result implies that the decision makers in the latter group have paid greater attention to the strengthening of sustainability progress capabilities and, hence, to the definition of a maturation route. © 2021
Detection of Sociolinguistic Features in Digital Social Networks for the Detection of Communities

E. Puertas, L.G. Moreno-Sandoval, J. Redondo, and 2 more authors

2021

Abs

The emergence of digital social networks has transformed society, social groups, and institutions in terms of the communication and expression of their opinions. Determining how language variations allow the detection of communities, together with the relevance of specific vocabulary (proposed by the National Council of Accreditation of Colombia (Consejo Nacional de Acreditación - CNA) to determine the quality evaluation parameters for universities in Colombia) in digital assemblages could lead to a better understanding of their dynamics and social foundations, thus resulting in better communication policies and intervention where necessary. The approach presented in this paper intends to determine what are the semantic spaces (sociolinguistic features) shared by social groups in digital social networks. It includes five layers based on Design Science Research, which are integrated with Natural Language Processing techniques (NLP), Computational Linguistics (CL), and Artificial Intelligence (AI). The approach is validated through a case study wherein the semantic values of a series of “Twitter” institutional accounts belonging to Colombian Universities are analyzed in terms of the 12 quality factors established by CNA. In addition, the topics and the sociolect used by different actors in the university communities are also analyzed. The current approach allows determining the sociolinguistic features of social groups in digital social networks. Its application allows detecting the words or concepts to which each actor of a social group (university) gives more importance in terms of vocabulary. © 2021, The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature.
Phonetic detection for Hate Speech Spreaders on Twitter

E. Puertas, and J.C. Martinez-Santos

2021

Abs

Nowadays, hate messages have become the object of study on social media. Efficient and effective detection of hate profiles requires various scientific disciplines, such as computational linguistics and sociology. Here, we illustrate how we used lexical and phonetic features to determine if the author spreads hate speech. This article presents a novel strategy for the characterization of the Twitter profile based on the generation of lexical and phonetic user features that serve as input to a set of classifiers. The results are part of our participation in the PAN 2021 in the CLEF in the task of Profiling Hate Speech Spreaders on Twitter. © 2021 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

2020

Assembly of polarity, emotion and user statistics for detection of fake profiles Notebook for PAN at CLEF 2020

L.G. Moreno-Sandoval, E. Puertas, A. Pomares-Quimbaya, and 1 more author

2020

Abs

The explosive growth of fake news on social networks has aroused great interest from researchers in different disciplines. To achieve efficient and effective detection of fake news requires scientific contributions from various disciplines, such as computational linguistics, artificial intelligence, and sociology. Here we illustrate how polarity, emotion, and user statistics can be used to detect fake profiles on Twitter’s social network. This paper presents a novel strategy for the characterization of the Twitter profile based on the generation of an assembly of polarity, emotion, and user statistics characteristics that serve as input to a set of classifiers. The results are part of our participation in the PAN 2020 in the CLEF in the task of Profiling Fake News Spreaders on Twitter. Copyright © 2020 for this paper by its authors.

2019

Celebrity profiling on twitter using sociolinguistic features notebook for PAN at CLEF 2019

L.G. Moreno-Sandoval, E. Puertas, F.M. Plaza-Del-Arco, and 3 more authors

2019

Abs

Social networks have been a revolutionary scenario for celebrities because they allow them to reach a wider audience with much higher frequency than using traditional means. These platforms enable them to improve or sometimes deteriorate, their careers through the construction of closer relationships with their fans and the acquisition of new ones. Indeed, networks have promoted the emergence of a new type of celebrities that exists only in the digital world. Being able to characterize the celebrities that are more active on social networks, such as Twitter, gives an enormous opportunity to identify what is their real level of fame, what is their relevance for an age group, or a specific gender or occupation. These facts may enrich decision making, especially in advertising and marketing. To achieve this aim, this paper presents a novel strategy for the characterization of celebrities profile on Twitter based on the generation of socio-linguistic features from their posts that serve as input to a set of classifiers. Specifically, we produced four classifiers that describe the level of fame, the gender, the birth date, and the possible occupation of a celebrity. We obtained the training and test data sets as part of our participation at PAN 2019 at CLEF. Results of each classifier are reported including the analysis of which features are more relevant, which classification techniques were more useful and which were the final precision and recall results. © 2019 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Bots and gender profiling on twitter using sociolinguistic features notebook for PAN at CLEF 2019

E. Puertas, L.G. Moreno-Sandoval, F.M. Plaza-Del-Arco, and 3 more authors

2019

Abs

Unfortunately, in social networks, software bots or just bots are becoming more and more common because malicious people have seen their usefulness to spread false messages, spread rumors and even manipulate public opinion. Even though the text generated by users in social networks is a rich source of information that can be used to identify different aspects of its authors, not being able to recognize which users are truly humans and which are not, is a big drawback. In this work, we describe the properties of our multilingual classification model submitted for PAN2019 that is able to recognize bots from humans, and females from males. This solution extracted 18 features from the user’s posts and applying a machine learning algorithm obtained good performance results. © 2019 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).

2018

An automatic approach to generate corpus in Spanish

E. Puertas, J.A. Alvarado-Valencia, L.G. Moreno-Sandoval, and 1 more author

2018

Abs

A corpus is an indispensable linguistic resource for any application of natural language processing. Some corpora have been created manually or semi-automatically for a specific domain. In this paper, we present an automatic approach to generate corpus from digital information sources such as Wikipedia and web pages. The information extracted by Wikipedia is done by delimiting the domain, using a propagation algorithm to determine the categories associated with a domain region and a set of seeds to delimit the search. The information extracted from the web pages is carried out efficiently, determining the patterns associated with the structure of each page with the purpose of defining the quality of the extraction. © Springer Nature Switzerland AG 2018.
Age classification from Spanish tweets the variable age analyzed by using linear classifiers

L.G. Moreno-Sandoval, J.F. Mendoza-Molina, E.A. Puertas, and 3 more authors

2018

Abs

Text classification or text categorization in social networks such as Twitter has taken great importance with the growth of applications of this process in diverse domains of society. Literature about text classifiers is significantly wide especially in languages such as English; however, this is not the case for age classification whose studies have been mainly focused on image recognition and analysis. This paper presents the results of testing linear classifiers performance in the task of identifying Twitter users age from their profile descriptions and tweets. For this purpose, a Spanish Lexicon of 45 words around the concept "cumpleaños" was created and the Gold Standard of 1541 users with age correctly identified was obtained. The experiments are presented with the description of the algorithms used to finally obtain the best seven models that permit to identify the user’s age with accuracy results between 66% and 69 %. Considering the information-retrieval layer, the new results showed that accuracy was increased from 69,09% to 72,96%. Copyright © 2018 by SCITEPRESS - Science and Technology Publications, Lda. All rights reserved.

2017

Constructor automático de modelos dominios sin corpus preexistente [Automatic constructor of domain models without pre-xisting corpus]

E.A. Puertas Del Castillo, J.A. Alvarado Valencia, and A.P. Quimbaya

2017

Abs

This project is about an automatic builder of domain models without pre-existing corpus. The constructor is based on techniques and methods for the construction of corpus from data extracted from digital media, through the use and development of software libraries that automate the phases of the process of building domains. This project is in phases of conceptual testing and component development. © 2017 Sociedad Espanola para el Procesamiento del Lenguaje Natural.