Created on 16th March 2024
•
We were provided a dataset from a Waveform generator, so we were asked to analyze the data and provide relevant insights. With 5000 instances, 21 attributes, and 3 classes, I gained visual insights and analyzed the data — classifying it into the three classes assigned.
• The dataset explored 5000 instances of 3 classes of waves, their noise levels, and the 21 associated attributes.
• We found that closely placed attributes correlated positively, while those placed away correlated negatively.
• The highest occurring value of an instance varied around 0.
• The waves were successfully classified into their respective classes (0, 1, 2) with 87.2% accuracy.
The most difficult hurdle was understanding the dataset. Since it was attributed to CART techniques, it took me a long time to understand that these were relevant waves to act as a test for CART testing and classification techniques.
Tracks Applied (1)
Technologies used