Institute of Formal and Applied Linguistics

Demo 1   -- Verb Patterns Classification

Preparing data for a supervised machine learning task
Manual annotation and basic data analysis 

• You will work with two English verbs, namely cry and enlarge.
• First, get familiar with the verbs. Look at their dictionary definitions.
  (The following excerpts are taken from the Collins Cobuild English Dictionary.)
  • collins.cry.pdf
  • collins.enlarge.pdf

 

Your task  --  manual annotation of a sample data

• You will be given 10+10 sentences with the given verbs
• You should recognize the pattern of verb usage
  • there are 3 predefined patterns for the verb cry 
    patterns.cry.pdf
  • there are 4 predefined patterns for the verb enlarge
    patterns.enlarge.pdf

 

• For each sentence you should assign a pattern that fits best the given sentence
  • if you do not think that no pattern matches the sentence, choose "u"
  • if you do not think that the given word is a verb, choose "x"

 

• Each student will take and fill just one of the following forms
  • Group A: A0, A1, A2, A3, A4, A5, A6, A7, A8, A9
  • Group B: B0, B1, B2, B3, B4, B5, B6, B7, B8, B9
  • Group C: C0, C1, C2, C3, C4, C5, C6, C7, C8, C9
  • Group D: D0, D1, D2, D3, D4, D5, D6, D7, D8, D9

Demo data

• Gold Standard
  • CRY cry-gs.csv
  • ENLARGE enlarge-gs.csv
• Automatic classifier (model F1)
  • CRY cry-F1.csv
  • ENLARGE enlarge-F1.csv 
• Students' annotations (60 examples, 4 groups, 2015)
  • CRY cry-A.csv, cry-B.csv, cry-C.csv, cry-D.csv
        -- Note (MH, Oct 10, 2015): File cry-B was wrong this morning, now it is already fixed.