Performing classification task on mobile price range.
Before moving to the training, we have made use of the SelectKBest method from sklearn, such as to select the 5 most impactful features to predict the target class price_range. The selected classes were 'battery_power', 'mobile_wt', 'px_height', 'px_width', 'ram', such that the range of each feature can be seen in the box plots below:
The models we have made use of are Logistic Regression and SVM, the results obtained for each model along with their confusion matrix can be seen below:
Logistic Regression testing acc: 98%
SVM testing acc: 96%
In order to use the gradio app interface, please clone the repo locally and use the following command
gradio app.py
The interface can be found on the following address http://127.0.0.1:7860