enhancing keyword extraction and categorization

[Asifdotexe]

Aim of the branch

To improve to accuracy but reducing the false positives and misclassification by the base Yake model.

[Asifdotexe]

The misclassification has been decreased comparatively, still there are cases where the algorithm misclassifies. One such case was for the keyword "heats up quickly"

We know that it is bad in context of tech such as smartphones, tablets, laptops etc. but the algorithm doesn't have the level of context about what we are reviewing, hence it misclassifies in cases like this.

Apart from this another thing to mention would be the runtime.

• The runtime has increased from the average of 23.21 seconds to 107.81 seconds.
• The current approach is roughly 4.64 times slower than the previous approach.