Development of a Sentiment Analysis Model for Evaluating Open Source Reviews on CCleaner
Autor(ka) práce:
Flores Trochez, Luis Diego
Typ práce:
Diploma thesis
Vedoucí práce:
Ziaei Nafchi, Majid
Oponenti práce:
Sudzina, František
Jazyk práce:
English
Abstrakt:
This study develops a machine learning-based sentiment analysis model to automate the evaluation of user reviews for CCleaner and two of its competitors. By collecting over 3,000 reviews from the Google Play Store and applying preprocessing, classification, and clustering techniques, the study compares the performance of Logistic Regression, Support Vector Machine (SVM), and Long Short-Term Memory (LSTM) models for sentiment classification. Results indicate that LSTM consistently outperforms traditional models across accuracy, precision, recall, and F1-score metrics. In addition, K-Means clustering reveals five dominant feedback themes, aiding product teams in pinpointing areas of user concern and satisfaction. The findings show the practical value of automated sentiment analysis in enhancing user experience, reducing manual effort, and informing decision makers of current user concerns.
Klíčová slova:
CCleaner; LSTM; Logistic Regression; Natural Language Processing (NLP); App Reviews; K-Means Clustering; Topic Modeling; Sentiment Analysis; Machine Learning; Support Vector Machine; Deep Learning
Název práce:
Development of a Sentiment Analysis Model for Evaluating Open Source Reviews on CCleaner
Autor(ka) práce:
Flores Trochez, Luis Diego
Typ práce:
Diplomová práce
Vedoucí práce:
Ziaei Nafchi, Majid
Oponenti práce:
Sudzina, František
Jazyk práce:
English
Abstrakt:
This study develops a machine learning-based sentiment analysis model to automate the evaluation of user reviews for CCleaner and two of its competitors. By collecting over 3,000 reviews from the Google Play Store and applying preprocessing, classification, and clustering techniques, the study compares the performance of Logistic Regression, Support Vector Machine (SVM), and Long Short-Term Memory (LSTM) models for sentiment classification. Results indicate that LSTM consistently outperforms traditional models across accuracy, precision, recall, and F1-score metrics. In addition, K-Means clustering reveals five dominant feedback themes, aiding product teams in pinpointing areas of user concern and satisfaction. The findings show the practical value of automated sentiment analysis in enhancing user experience, reducing manual effort, and informing decision makers of current user concerns.
Klíčová slova:
Sentiment Analysis; Machine Learning; Support Vector Machine; Logistic Regression; App Reviews; Natural Language Processing (NLP); CCleaner; Deep Learning; Topic Modeling; K-Means Clustering; LSTM