अभ्यास

Good data vs. bad data

Here's an excerpt from a training set that labels the entity type TOURIST_DESTINATION in traveler reviews.

निर्देश 1/2

undefined XP

1

2

प्रश्न

Why is this data and label scheme problematic?

संभावित उत्तर

Whether a place is a tourist destination is a subjective judgement and not a definitive category. It will be very difficult for the entity recognizer to learn.

"Paris" and "Arkansas" should also be labelled as tourist destinations for consistency. Otherwise, the model will be confused.

Rare out-of-vocabulary words like the misspelled "amsterdem" shouldn't be labelled as entities.

.css-6su6fj{-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;}अभ्यास

निर्देश 1/2

प्रश्न

संभावित उत्तर

अभ्यास