search

guillaumegenthial / tf_ner

Simple and Efficient Tensorflow implementations of NER models with tf.estimator and tf.data
Apache License 2.0
924 stars 275 forks source link

What does it mean ? - Results score #15

Closed rodmandi closed 5 years ago

rodmandi commented 6 years ago

seleccion_024

Well as you can see, we have in "Francisco" I-LOC B-LOC What does it mean? What is the meaning of I-LOC and B-LOC? i don't know if the meaning of I-LOC is that tag was the correct tag, and B-LOC was the the predicted tag, or maybe another example Moscow B-LOC B-LOC.

Thank you!

guillaumegenthial commented 6 years ago

Hi @rodmandi,

The first tag is the true tag (from the dummy dataset), the second tag is the prediction. I assume you trained the model just on this dummy dataset, which would result in a crappy model (almost perfectly overfitted on the dummy dataset).

So in Francisco I-LOC B-LOC it means that the model made a mistake by predicting B-LOC instead of I-LOC.

Hope that helps

bhushanbrb commented 6 years ago

@rodmandi Just to add more In case you do not know this may help ... B-I-O is Begin Inside Outside Input data pattern. Explanation from @guillaumegenthial blog In the CoNLL2003 task, the entities are LOC, PER, ORG and MISC for locations, persons, orgnizations and miscellaneous. The no-entity tag is O. Because some entities (like New York) have multiple words, we use a tagging scheme to distinguish between the beginning (tag B-...), or the inside of an entity (tag I-...). Other tagging schemes exist (IOBES, etc). However, if we just pause for a sec and think about it in an abstract manner, we just need a system that assigns a class (a number corresponding to a tag) to each word in a sentence. --> What is the meaning of I-LOC and B-LOC? Word from sequence | TrueTag | PredictedTag