Learning Pathways and Students Performance: A Dynamic Complex System.

In this study, learning pathways are modelled by networks constructed from the log data of student-LMS interactions. These networks capture the sequence of reviewing the learning materials by the students enrolled in a given course. In previous research, the networks of successful students showed a fractal property; meanwhile, the networks of students who failed showed an exponential pattern. This research aims to provide empirical evidence that students' learning pathways have the properties of emergence and non-additivity from a macro level; meanwhile, equifinality (same end of learning process but different learning pathways) is presented at a micro level. Furthermore, the learning pathways of 422 students enrolled in a blended course are classified according to learning performance. These individual learning pathways are modelled by networks from which the relevant learning activities (nodes) are extracted in a sequence by a fractal-based method. The fractal method reduces the number of nodes to be considered relevant. A deep learning network classifies these sequences of each student into passed or failed. The results show that the accuracy of the prediction of the learning performance was 94%, the area under the receiver operating characteristic curve was 97%, and the Matthews correlation was 88%, showing that deep learning networks can model equifinality in complex systems.

Identification of Relevant Protein Interactions with Partial Knowledge: A Complex Network and Deep Learning Approach.

Protein-protein interactions (PPIs) are the basis for understanding most cellular events in biological systems. Several experimental methods, e.g., biochemical, molecular, and genetic methods, have been used to identify protein-protein associations. However, some of them, such as mass spectrometry, are time-consuming and expensive. Machine learning (ML) techniques have been widely used to characterize PPIs, increasing the number of proteins analyzed simultaneously and optimizing time and resources for identifying and predicting protein-protein functional linkages. Previous ML approaches have focused on well-known networks or specific targets but not on identifying relevant proteins with partial or null knowledge of the interaction networks. The proposed approach aims to generate a relevant protein sequence based on bidirectional Long-Short Term Memory (LSTM) with partial knowledge of interactions. The general framework comprises conducting a scale-free and fractal complex network analysis. The outcome of these analyses is then used to fine-tune the fractal method for the vital protein extraction of PPI networks. The results show that several PPI networks are self-similar or fractal, but that both features cannot coexist. The generated protein sequences (by the bidirectional LSTM) also contain an average of 39.5% of proteins in the original sequence. The average length of the generated sequences was 17% of the original one. Finally, 95% of the generated sequences were true.

An Entropy-Based Measure of Complexity: An Application in Lung-Damage.

The computed tomography (CT) chest is a tool for diagnostic tests and the early evaluation of lung infections, pulmonary interstitial damage, and complications caused by common pneumonia and COVID-19. Additionally, computer-aided diagnostic systems and methods based on entropy, fractality, and deep learning have been implemented to analyse lung CT images. This article aims to introduce an Entropy-based Measure of Complexity (EMC). In addition, derived from EMC, a Lung Damage Measure (LDM) is introduced to show a medical application. CT scans of 486 healthy subjects, 263 diagnosed with COVID-19, and 329 with pneumonia were analysed using the LDM. The statistical analysis shows a significant difference in LDM between healthy subjects and those suffering from COVID-19 and common pneumonia. The LDM of common pneumonia was the highest, followed by COVID-19 and healthy subjects. Furthermore, LDM increased as much as clinical classification and CO-RADS scores. Thus, LDM is a measure that could be used to determine or confirm the scored severity. On the other hand, the d-summable information model best fits the information obtained by the covering of the CT; thus, it can be the cornerstone for formulating a fractional LDM.

Fractional Online Learning Rate: Influence of Psychological Factors on Learning Acquisition.

The quantification of learning acquisition in a blended and online course is still slightly explored from the complex systems lens. The fractional online learning rate (fOLR) using fractional integrals is introduced. The notion of fOLR is based on the nonlinearity of the individual students learning pathway network, built from Learning Management System log files. Several learning pathway networks from students that pass or fail the course were constructed. The Akaike information criterion shows that the minimum number of boxes to cover these networks follow a power-law model. Further analysis shows that the fOLR model and its parameters were significantly compared with the online learning rate model. Thus, the fOLR was computing power and delayed power models, inspired by the "law of practice." The results show that the fractional definition is a better model and has a nonlinear relationship with the overall grade. Also, engagement and disengagement mould the fOLR curve. It means that the student's performance is affected by the engagement, and it is necessary that they are encouraged to pay more effort and attention to the learning activities, and those activities need to be designed to be fun and pleasant to improve the learning achievements.

Factores que cuantifican la percepción de calidad en el servicio al cliente en un restaurante mexicano / Factors that quantify the perception of quality in customer service in a Mexican restaurant

Resumen Una de las industrias más destacadas de la economía mexicana es la restaurantera. Su importancia, debido a su número de empresas, creación de empleos y emprendimientos, ha ocasionado que se genere un alto índice de competitividad. Esto provoca que se busquen estrategias para mejorar la calidad del servicio que ofrecen, con el propósito de retener y atraer clientes. El objetivo de este trabajo fue identificar los factores que conforman la percepción de la calidad en el servicio en un restaurante mexicano. Para ello, se utilizó el instrumento DINESERV, mediante un enfoque cuantitativo y un análisis factorial confirmatorio. Los resultados mostraron que el instrumento DINESERV es válido para restaurantes mexicanos. Asimismo, se detectaron los factores que integran el servicio al cliente, enfatizando los aspectos de tangibilidad, confiabilidad, respuesta y empatía. Características como personal competente y con experiencia, tener siempre presente los intereses del cliente y la apariencia de la vestimenta y limpieza del personal de servicio son elementos clave para que el restaurante genere mayor satisfacción en sus clientes.

Abstract One of the most prominent industries in the Mexican economy is the restaurant industry. Its importance, due to the number of companies, job creation and business ventures, has caused a high competitiveness index to be generated. This causes the search of strategies to be sought to improve the quality of the service they offer, in order to retain and attract customers. The objective of this work was to identify the factors that comprise service quality perception in a Mexican restaurant. For that purpose, we employed the DINESERV instrument, through a quantitative approach and a confirmatory factor analysis. Results showed that the DINESERV instrument is valid for Mexican restaurants. Likewise, the factors that make up customer service were identified, emphasizing the aspects such as tangibility, reliability, response and empathy. Factors such as competent and experienced staff, always keeping in mind the interests of the client, the appearance of the service personnel´s clothing and cleanliness are key elements for the restaurant to generate greater satisfaction in its customers.

A Two-Parameter Fractional Tsallis Decision Tree.

Decision trees are decision support data mining tools that create, as the name suggests, a tree-like model. The classical C4.5 decision tree, based on the Shannon entropy, is a simple algorithm to calculate the gain ratio and then split the attributes based on this entropy measure. Tsallis and Renyi entropies (instead of Shannon) can be employed to generate a decision tree with better results. In practice, the entropic index parameter of these entropies is tuned to outperform the classical decision trees. However, this process is carried out by testing a range of values for a given database, which is time-consuming and unfeasible for massive data. This paper introduces a decision tree based on a two-parameter fractional Tsallis entropy. We propose a constructionist approach to the representation of databases as complex networks that enable us an efficient computation of the parameters of this entropy using the box-covering algorithm and renormalization of the complex network. The experimental results support the conclusion that the two-parameter fractional Tsallis entropy is a more sensitive measure than parametric Renyi, Tsallis, and Gini index precedents for a decision tree classifier.

Classification of Literary Works: Fractality and Complexity of the Narrative, Essay, and Research Article.

A complex network as an abstraction of a language system has attracted much attention during the last decade. Linguistic typological research using quantitative measures is a current research topic based on the complex network approach. This research aims at showing the node degree, betweenness, shortest path length, clustering coefficient, and nearest neighbourhoods' degree, as well as more complex measures such as: the fractal dimension, the complexity of a given network, the Area Under Box-covering, and the Area Under the Robustness Curve. The literary works of Mexican writers were classify according to their genre. Precisely 87% of the full word co-occurrence networks were classified as a fractal. Also, empirical evidence is presented that supports the conjecture that lemmatisation of the original text is a renormalisation process of the networks that preserve their fractal property and reveal stylistic attributes by genre.

Estudio comparativo de técnicas de analítica del aprendizaje para predecir el rendimiento académico de los estudiantes de educación superior / Comparative study of learning analytics techniques to predict student academic performance in higher education

Resumen La deserción escolar involucra diversos factores, entre ellos, el compromiso del estudiante, a través del cual se puede predecir su éxito en la escuela. Ese compromiso tiene varios componentes, tales como conductual, emocional y cognitivo. La motivación y el compromiso están fuertemente relacionadas, ya que la primera es un precursor del compromiso. El objetivo de este estudio fue comparar la eficacia de la regresión lineal contra dos técnicas de minería de datos para predecir el rendimiento académico de los estudiantes en la educación superior. Se hizo un estudio transversal explicativo en el que se encuestó a 222 estudiantes universitarios de una institución pública de la Ciudad de México. Se realizó un análisis de regresión lineal jerárquico (RL) y de técnicas de analítica del aprendizaje, como redes neuronales (RN) y máquinas de vector soporte (SVM). Para evaluar la exactitud de las técnicas de analítica del aprendizaje se realizó un análisis de varianza (ANOVA). Se compararon las técnicas de analítica del aprendizaje y de regresión lineal usando la validación cruzada. Los resultados mostraron que el compromiso conductual y la autoeficacia tuvieron efectos positivos en el desempeño del estudiante, mientras que la pasividad mostró un efecto negativo. Asimismo, las técnicas de RL y de SVM pronosticaron igualmente el desempeño académico de los estudiantes. La RL tuvo la ventaja de producir un modelo simple y de fácil interpretación. Por el contrario, la técnica de SVM generó un modelo más complejo, aunque, si el modelo tuviese como objetivo el pronóstico del desempeño, la técnica SVM sería la más adecuada, ya que no requiere la verificación de ningún supuesto estadístico.

Abstract The issue of school dropout involves factors such as students' engagement that can predict his or her success in school. It has been shown that student engagement has three components: behavioral, emotional and cognitive. Motivation and engagement are strongly related since the former is a precursor of engagement. The aim of this study was to compare the efficiency of linear regression against two data mining techniques to predict the students' academic performance in higher education. A descriptive cross-sectional study was carried out with 222 students from a public higher education institution in Mexico city. An analysis of hiererchical linear regression (LR) and learning analytics techniques such as neural networks (NN) and support vector machine (SVM) was conducted. To assess the accuracy of the learning analytics techniques, an analysis of variance (ANOVA) was carried out. The techniques were compared using cross validation. The results showed that behavioral engagement and self-efficacy had positive effects on student achievements, while passivity showed a negative effect. Likewise, the LR and SVM techniques had the same performance on predicting students' achievements. The LR has the advantage of producing a simple and easy model. On the contrary, the SVM technique generates a more complex model. Although, if the model were aimed to forecast the performance, the SVM technique would be the most appropriate, since it does not require to verify any statistical assumption.

Fractional information dimensions of complex networks.

In this article, new information dimensions of complex networks are introduced underpinned by fractional order entropies proposed in the literature. This fractional approach of the concept of information dimension is applied to several real and synthetic complex networks, and the achieved results are analyzed and compared with the corresponding ones obtained using classic information dimension based on the Shannon entropy. In addition, we have investigated an extensive classification of the treated complex networks in correspondence with the fractional information dimensions.