Page 171 - Emerging Trends and Innovations in Web-Based Applications and Technologies

International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
Precision = True Positives
True Positives + False Positives
 Recall: Recall measures how often the classifier correctly identifies positive instances from all actual positive instances. It
is calculated as:
Recall = True Positives
True Positives + False Negatives
 F1 Score: The F1 score is the harmonic mean of precision and recall, providing a balanced measure of the classifier's
performance. It is calculated as:
F1-Score = (2 * Precision * Recall)
(Precision + Recall)
By using these metrics, Quick Mart can ensure that the CNN model effectively classifies second-hand products and delivers an
enhanced user experience in the marketplace. These performance evaluations help in continuously refining the model and
adjusting the system to improve accuracy and minimize errors in product categorization.
VI. RESULT ANALYSIS
The experiments were conducted using a computer with an Intel Core i5 CPU and 4GB of RAM, with Jupyter Notebook
facilitating the development and training of smart solutions for Quick Mart’s second-hand marketplace. The experimental
outcomes demonstrate a significant improvement in the marketplace's operational accuracy, with an efficiency of 92.14% for
the proposed solution. This system effectively identifies and categorizes items, streamlining the second-hand trading process.

Figure 4: Model Training and Validation Accuracy
Figure 5 shows the performance of the proposed custom-designed model, where the blue and orange curves represent
validation and training accuracy, respectively. The x-axis denotes the number of epochs, while the y-axis shows the percentage
accuracy. The plot indicates that as the number of epochs increases, the training accuracy remains consistently high. However,
validation accuracy is slightly lower in comparison, which is typical of model overfitting. Despite this, the solution
demonstrates a solid level of performance with significant accuracy across various items in the marketplace.

Figure 5: Model Training and Validation Loss

IJTSRD | Special Issue on Emerging Trends and Innovations in Web-Based Applications and Technologies Page 161