http://hdl.handle.net/11422/52 2026-04-07T01:01:57Z 2026-04-07T01:01:57Z Dessabato, Karolayne Pereira http://hdl.handle.net/11422/27330 2025-10-08T03:00:09Z 2025-01-01T00:00:00Z

Title: Machine learning methods in music emotion recognition Author(s)/Inventor(s): Dessabato, Karolayne Pereira Advisor: Carvalho, Hugo Tremonte de Abstract: Music Emotion Recognition (MER), an area within Musical Information Retrieval (MIR), studies the emotions evoked in listeners by music. We address MER as a regression task, with the objective of predicting the emotional content of music (encoded in arousal and valence) from acoustic features extracted from the waveform. We apply an interpretable machine learning technique, investigating the role of these features in predicting the target variables. Initially, a random forest model is trained on the DEAM dataset (MediaEval Database for Emotional Analysis of Music). Then, we use the concept of Shapley values to interpret the role of each variable in the predictions made by this model. Finally, we extract the most significant features from the DEAM dataset to predict arousal and valence, thus enhancing the interpretability of the model employed. Additionally, we explore a dynamic linear model approach to gain further insights into the relationships between features and response variables. This method allows for a potentially “less black-box” and more interpretable representation of the problem. Principal Component Analysis (PCA) is also utilized to analyze the structure of features in the dataset, providing a more comprehensive understanding of the key variables influencing MER predictions. By integrating these approaches, we aim to enhance both the predictive performance and interpretability of the models, offering meaningful insights into the most relevant features that drive emotional responses in music. Publisher: Universidade Federal do Rio de Janeiro Type: Dissertação

2025-01-01T00:00:00Z Bastos, Ismael Sampaio http://hdl.handle.net/11422/27329 2025-10-08T03:00:09Z 2025-01-01T00:00:00Z

Title: Algoritmos para multi-armed bandits: teoria e aplicação à precificação dinâmica Author(s)/Inventor(s): Bastos, Ismael Sampaio Advisor: Iacobelli, Giulio Abstract: This work addresses the problem of sequential decision-making, focusing specifically on the multiarmed bandit (MAB) framework. In its classical formulation, the MAB problem involves an agent facing a row of slot machines (bandits), with a limited number of pulls (arms) available. The agent’s goal is to determine a sequence of actions that maximizes the total reward. The core challenge lies in balancing the trade-off between choosing the action that currently appears to yield the highest reward and exploring lesser-known alternatives (a dilemma known as exploration versus exploitation). In this study, we explore several algorithms designed to support decision-making within the multiarmed bandit setting. We also examine an application of this theory to the problem of dynamic pricing, i.e., determining optimal selling prices for products and services. In this context, the seller takes the role of the agent who aims to sell a product by selecting from a finite set of possible prices, without prior knowledge of demand or consumer behavior. The seller must therefore adopt a strategy that enables the identification of the optimal price over time. Publisher: Universidade Federal do Rio de Janeiro Type: Dissertação

2025-01-01T00:00:00Z Argote Osorio, Juan Pablo http://hdl.handle.net/11422/27327 2025-10-08T03:00:09Z 2025-02-24T00:00:00Z

Title: Alocação latente de Dirichlet para modelagem de tópicos em dissertações de mestrado em estatística e áreas correlatas no Brasil Author(s)/Inventor(s): Argote Osorio, Juan Pablo Advisor: Pagani Zanini, Carlos Tadeu Abstract: This master’s thesis addresses the topic modeling of master’s theses in statistics and related areas in Brazil, through Latent Dirichlet Allocation models. The main objective of the work is to infer the latent topics covered in these theses. First, the construction of a corpus of documents is discussed and presented, composed of the most recent theses from different Higher Education Institutions in Brazil, manually extracted from the web pages of each of the analyzed master’s programs. The inferential procedure adopted for the Latent Dirichlet Allocation model consists of Markov chain Monte Carlo methods and variational inference. Different methods for choosing the number of topics are also discussed, including information criteria such as Akaike, Bayesian, Deviance, Watanabe-Akaike, and metrics based on the coherence of the inferred latent topics. The adopted methodology provides an in-depth understanding of the predominant topics in this corpus. Publisher: Universidade Federal do Rio de Janeiro Type: Dissertação

2025-02-24T00:00:00Z Arce, Luiza do Bonfim http://hdl.handle.net/11422/27321 2025-10-08T03:00:08Z 2025-02-17T00:00:00Z

Title: Produção de indicadores do mercado de trabalho com modelos de estimação em pequenas áreas para dados composicionais Author(s)/Inventor(s): Arce, Luiza do Bonfim Advisor: Gonçalves, Kelly Cristina Mota Abstract: The demand for increasingly disaggregated indicators is rising, mainly as a result of the 2030 Agenda. Indicators can describe important characteristics of the population and provide a basis for public policies, allocation of resources, among other uses. Therefore, ensuring that their estimates are representative and reliable is essential. However, when an indicator is calculated based on data from sample surveys, its precision may sometimes fall below the threshold suitable for dissemination to certain population groups. This is the case for the proportion of unemployed people and the unemployment rate for municipal strata, both calculated using data from the Continuous National Household Sample Survey (PNADC). In the third quarter of 2023, for both indicators, some strata had estimates that were either not very precise or imprecise, according to IBGE classification. In this context, Small Area Estimation can be used, which consists of employing models that allow the inclusion of auxiliary data about population groups to assist in the estimation process. Thus, Dirichlet and Generalized Dirichlet models, which are standard models for compositional data, were proposed as alternatives for estimating labor market indicators, such as the unemployment rate, the proportion of unemployed, employed, and people outside the labor force, with the aim of improving the precision of the unemployment rate and the proportion of unemployed people. It was expected that, by jointly estimating the proportions of interest, there would be a significant increase in the precision of the three proportions and the unemployment rate, which can be derived from the proportions. Publisher: Universidade Federal do Rio de Janeiro Type: Dissertação

2025-02-17T00:00:00Z