This is an hands-on workshop for total beginners in Natural Language Processing who are already proficient with python.

Potential audience: Whoever is interested in experimenting with this fascinating domain of NLP, whether you're a pro data scientist with no experience in NLP or not a dev guy (PM, bizdev, ...) but can code in Python.

Please register here: http://yarok-hok.com/events/nlp-workshop/

We would cover text classification in python, using sklearn, spacy, nltk and pandas, via a challenge (and competition) over real data.

This event is hands-on, All attendees MUST bring:

1. A laptop with python 3.6 installed (preferably anaconda for python3)
2. Download Data: http://goren.ml/pdnlp
3. Clone repository: https://github.com/urigoren/nlp_classification

Agenda (Short):

16:30-17:00 - Gathering
17:00 - 18:30 - Natural language processing introduction (3 short lectures)
18:30 Hands-on workshop & Competition Start
21:00 Competition End, 'and the winner is...'

Agenda (Long):

16:30-17:00 - Gathering
17:00 - 17:15 - Natural language processing intro

• A brief overview of NLP tasks
• Supervised tasks (Named entity recognition, Sentiment analysis, classification)
• Unsuperised tasks (Text generation, machine translation, topic modeling)
• Why is NLP harder in Hebrew
• overview of the data set
  17:15 - 18:00 From Textual documents to vectors
• Preprocessing (String manipulations and Defining word boundaries and tokens)
• Word stems / lemmas
• Identifying phrases
• Generating custom vocabularies
• Transforming a document to a vector
• One hot word encodings
• Word Vectors (word2vec, glove)
• Combining word vectors in document vectors with the Bag Of Words assumption
  18:00 - 18:30 Modelling in depth
• Logistic regression for document classification
• Naive Bayes modeling
• Model evaluation
• Training and testing
• Metrics
  18:30 Hands-on workshop and Competition Start
  20:30 Competition End