Bibliographic citations
Flores, F., (2024). Clasificación de datos textuales provenientes de un streaming aplicando el método de representación de texto TF-IDF en una Regresión Logística [Universidad Nacional Agraria La Molina]. https://hdl.handle.net/20.500.12996/6638
Flores, F., Clasificación de datos textuales provenientes de un streaming aplicando el método de representación de texto TF-IDF en una Regresión Logística []. PE: Universidad Nacional Agraria La Molina; 2024. https://hdl.handle.net/20.500.12996/6638
@misc{renati/1114986,
title = "Clasificación de datos textuales provenientes de un streaming aplicando el método de representación de texto TF-IDF en una Regresión Logística",
author = "Flores Cáceres, Fiorella Alexandra",
publisher = "Universidad Nacional Agraria La Molina",
year = "2024"
}
The purpose of this research work was to implement a logistic regression model using transformed textual data using the TF-IDF text representation method, with the aim of classifying teacher comments in guidance streamings on the “Aprendo en Casa” strategy carried out by the Ministry of Education. The analysis procedure was divided into data preprocessing, exploratory data analysis, application of the TF-IDF text representation method, model estimation and evaluation; and classification of new comments. For the preprocessing stage, the textual data of the comments were cleaned and standardized; while in the exploratory analysis, descriptive indicators of the comments of each category were obtained using n-grams. In the application of the TF-IDF text representation method, the document-term matrix was created from the training sample and the Chi-Square test was used for variable selection. In the estimation of the classification model, the final model adjusted with the training data from the document-term matrix was obtained. To evaluate the model, the TF-IDF method was applied to the test sample, in order to obtain its document-term matrix to perform the classification and find the results of the evaluation metrics, where an accuracy of 0.81 was achieved. Subsequently, the classification model was evaluated using the K-Fold Cross-Validation method and new comments were classified. Based on the results of this research, it is concluded that the implementation of the developed model is adequate.
This item is licensed under a Creative Commons License