Questions: Predictability of Distributional Semantics in WF

Hints before reading the paper:

• you can skip most of chapter 4, specifically from page 1291 to the end of the chapter;
• Distributional_Hypothesis says that words that occur in the same contexts tend to have similar meanings. The term distribution in the field of Distributional semantics is understood as a distribution of contexts of a given word in a corpus, where the context is usually defined by a window of N neighboring words.

Questions

1. Determine base words and derivational patterns (similarly as in Table 1) for the following words: softness (noun), undemocratic (adjective), text (verb), donate (verb). Try to propose another example for each pattern.

2. Which method did the authors use for vector representation of words in the prediction models? How many prediction models did the authors train?

3. Calculate the Mean reciprocal rank for the data below. Hint: see Mean reciprocal rank on wikipedia.

word	ranked results
softness	1: softly, 2: soft, 3: soften
undemocratic	1: democratic, 2: democracy, 3: democrat
donate	1: donator, 2: donatress, 3: donater