Abstract:
Algorithms for machine learning (ML) play a crucial role in various industries such as e-commerce, banking, education, and industry. The performance of ML algorithms depends on factors like the dataset and processing stages. To achieve accurate results, it is essential to select the appropriate algorithm and employ effective preprocessing and postprocessing techniques. This study focuses on predicting the pricing class of mobile phones and compares several ML techniques including Random Forest Classifier, Logistic Regression Classifier, Decision Tree Classifier, Linear Discriminant Analysis, K-Nearest Neighbor Classifier, and SVC. The evaluation is performed using the "Mobile Price Classification" dataset from Kaggle.com. Initially, the dataset is checked for missing entries, followed by scaling to enhance data usability for ML algorithms. Feature selection techniques are employed to identify relevant features and reduce computational costs by minimizing inputs. Finally, the classification algorithm parameters are tuned to improve system accuracy. The findings demonstrate that the ANOVA f-test feature selection method is practical for this dataset, as it achieves satisfactory accuracy with a minimal number of features. Moreover, among the compared models, the SVC classifier exhibits the highest test accuracy.
This final year engineering project is based on an IEEE paper and is considered one of the best projects for computer science students. As the best IEEE project center in Bangalore, we offer complete documentation support, fully implemented hardware/software in the students' environment, and organized classes to facilitate learning and project development.