Training and Prediction Data Discrepancies: Challenges of Text Classification with Noisy, Historical Data.

R. Andrew Kreek Emilia Apostolova

Published in: NUT@EMNLP (2018)

Keyphrases

historical data
text classification
data mining techniques
predictive model
stream data
data sets
data sources
data points
data processing
data quality
data cleaning
labeled data
training samples
real world
knowledge discovery
training data
noisy data
learning algorithm
neural network