International Journal for Applied Information Management

Exploring Thematic Travel Preferences of Global Cities through Agglomerative Hierarchical Clustering for Enhanced Travel Recommendations

Soeltan Abdul Ghaffar — 2025-11-21

This study explores the application of Agglomerative Hierarchical Clustering (AHC) to categorize global cities based on thematic travel preferences, aiming to enhance personalized travel recommendations. The dataset used contains travel information for 560 cities worldwide, including thematic ratings across nine categories: culture, adventure, nature, beaches, nightlife, cuisine, wellness, urban, and seclusion, along with climate data and city descriptions. Feature engineering was performed to calculate an overall rating for each city by averaging its thematic scores, and to compute an average annual temperature from monthly climate data. The primary objective of this research was to use AHC to group cities into distinct clusters based on these thematic ratings. The analysis revealed six clusters, each representing different types of travel experiences. Cluster 1 consists of urban cultural hubs with high ratings for culture, cuisine, and urban experiences, while Cluster 2 features cities with a balance of cultural and culinary experiences alongside moderate natural and nightlife attractions. Cluster 3 represents remote, nature-focused cities with high ratings for seclusion and nature. Cluster 4 includes cities renowned for their beaches, nature, and cuisine, while Cluster 5 groups cities that emphasize adventure, nature, and seclusion. Cluster 6 is made up of destinations with a focus on nature, adventure, and seclusion, offering a balance between outdoor activities and tranquility. These findings offer a deeper understanding of the diversity in global city offerings and can significantly improve the effectiveness of travel recommendation systems by aligning cities with users' thematic preferences. By categorizing cities into meaningful clusters, personalized travel suggestions can be made based on users’ specific interests, such as cultural exploration, adventure, or nature. This research lays the groundwork for future studies to incorporate additional data sources and explore alternative clustering techniques for even more refined travel recommendations. The practical applications of this research can enhance real-world travel recommendation platforms, making them more tailored and relevant to individual user preferences

Leveraging TF-IDF and Random Forest to Uncover Genre Patterns in Google Books Metadata

Nadya Awalia Putri — 2025-11-21

This paper presents a machine learning-based approach for classifying books into genres using their descriptions. We employed a Random Forest classifier combined with Term Frequency-Inverse Document Frequency (TF-IDF) to convert text descriptions into numerical features, enabling the classification of books into six genres: Fiction, Literary Criticism, Education, Social Science, Biography & Autobiography, and Unknown Genre. The model was trained and evaluated on a dataset sourced from Google Books, which was preprocessed to remove missing data and clean the text descriptions by eliminating punctuation, numbers, and stopwords. We performed 5-fold cross-validation to assess the model's performance, which resulted in an average cross-validation accuracy of 64.22%. The final model achieved an accuracy of 62.71% on the test set, with the highest recall observed in the "Fiction" genre. The results indicated that the Random Forest classifier was particularly effective in classifying well-represented genres like "Fiction" and "Unknown Genre." However, genres with fewer samples, such as "Social Science" and "Biography & Autobiography," showed poor performance, highlighting the challenges posed by class imbalance and data sparsity. A confusion matrix and classification report revealed these discrepancies, with certain genres being misclassified more often than others. This research demonstrates the feasibility of using machine learning for automated book genre classification, offering significant potential for enhancing book recommendation systems and improving user experience. Despite its promising results, the study's limitations, including data sparsity and genre imbalance, suggest that further work is needed to refine the model. Future research could explore the use of deep learning techniques and the expansion of the dataset to address these issues and improve genre classification accuracy. The potential for automated genre classification in real-world applications, such as book categorization and personalized recommendations, presents an exciting direction for the book industry.

Classifying Vehicle Categories Based on Technical Specifications Using Random Forest and SMOTE for Data Augmentation

Dwi Sugianto — 2025-11-21

This study investigates the application of machine learning for classifying vehicles based on their technical specifications using the Random Forest algorithm. The objective was to create a robust classification model capable of categorizing vehicles into six distinct classes: Hybrid, SUV, Sedan, Sports, Truck, and Wagon. The analysis was conducted using a comprehensive dataset that included features such as engine size, horsepower, weight, and fuel efficiency, along with the target variable, vehicle class. To address the issue of class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was applied to balance the training data. The results showed that the model performed particularly well in classifying Sedans, achieving a perfect recall and high F1-score, while struggling with underrepresented classes like Hybrid and Wagon. Despite applying SMOTE, the model’s performance for minority classes remained suboptimal, highlighting the challenges associated with highly imbalanced datasets. The study contributes to the field of vehicle classification by demonstrating the use of Random Forest for such tasks and providing insights into the challenges posed by imbalanced class distributions. The findings underscore the importance of feature selection, especially regarding numerical attributes such as horsepower and engine size, in improving classification accuracy. However, the study also identified limitations, including potential biases in the dataset and the difficulty in improving performance for minority vehicle classes. Future research should explore alternative algorithms like XGBoost or deep learning models, and consider expanding the dataset to include more diverse vehicle types. The practical implications of this work are significant for vehicle market segmentation, offering valuable insights for manufacturers, dealerships, and analysts seeking to optimize vehicle classification and improve market targeting strategies.

Evaluating the Performance of Random Forest Algorithm in Classifying Property Sale Amount Categories in Real Estate Data

Les Endahti — 2025-11-21

This study explores the use of machine learning algorithms to classify property sale categories in real estate data, focusing on the performance of the Random Forest algorithm. The dataset, comprising over one million records of property sales from 2001 to 2022, includes features such as sale amount, assessed value, sales ratio, property type, and residential type. The primary objective is to determine which algorithm better predicts property sale categories and to assess how these predictions can aid in market segmentation and property valuation. After preprocessing the data by removing irrelevant columns and handling missing values, we applied the Random Forest classifier to predict five key property types: 'Single Family', 'Residential', 'Condo', 'Two Family', and 'Three Family'. The model achieved an accuracy of 82.98%, with high recall for categories like 'Single Family' and 'Condo', but struggled with 'Residential', which displayed a lower recall due to its diverse nature. The findings suggest that the Random Forest algorithm performs well in predicting certain property types, but improvements are needed for categories with more variation. The study highlights the importance of selecting relevant features such as sale amount and assessed value, which were found to be the most influential in determining property type. Real estate professionals can leverage these machine learning models for more accurate market segmentation, leading to better pricing and marketing strategies. However, the study also acknowledges limitations, such as the complexity of the 'Residential' category and potential data imbalance. Future research could focus on incorporating additional features, such as location-specific data or detailed property descriptions, and testing alternative algorithms to further enhance classification accuracy.

Exploring Football Player Salary Prediction Using Random Forest: Leveraging Player Demographics and Team Associations

Riyadh Abdulhadi M Aljohani — 2025-11-21

This paper explores the prediction of football player salaries using a Random Forest Regressor model, leveraging player demographics and team associations as key features. The dataset consists of 684 football players, including variables such as age, nationality, position, team, weekly salary, and annual salary. The study applies exploratory data analysis (EDA) to understand the distribution of these features and identify patterns within the dataset. Data preprocessing involves handling missing values, one-hot encoding categorical variables, and splitting the dataset into training and testing sets. The Random Forest model is trained on the preprocessed data, and its performance is evaluated using common regression metrics, including R-squared (R²), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). The results show that the model explains approximately 48.5% of the variance in player salaries, with an MAE of £1.92 million and an RMSE of £2.82 million. Key predictors of salary include player age, position, nationality, and team. The analysis of feature importance reveals that categorical variables such as Nation and Team have a significant impact on salary predictions. However, the model's performance is constrained by the lack of more granular data, such as player performance metrics or external economic factors. This research provides valuable insights for football team management, helping teams understand which factors contribute to salary setting and enabling more informed decisions in player recruitment and contract negotiations. It also highlights the potential for sponsorships to target players based on these predictive attributes. Future work could explore the integration of more advanced machine learning techniques and additional player data to improve predictive accuracy and model robustness.