5.) compute PPMI (information, sugar) using data in below table
computer
2
data
result
pie
8
9
442
cherry strawberry digital information
1
60
1670
3325
1683
3982
85
5
378
5
count (context)
4997
5673
473
512
sugar
25
19
4
13
61
count(W)
486
80
3447
7703
11716
To compute the PPMI (Pointwise Mutual Information) between "information" and "sugar," we need to calculate the following values:
1. Count("information", "sugar"): This is the number of times "information" and "sugar" co-occur in the data. In this case, it is 5.
2. Count("information"): This is the total number of occurrences of the word "information" in the data. In this case, it is 473.
3. Count("sugar"): This is the total number of occurrences of the word "sugar" in the data. In this case, it is 61.
4. Total count (W): This is the total number of words in the data. In this case, it is 11716.
Now, let's calculate the PPMI.
PPMI("information", "sugar") = log2( (Count("information", "sugar") / Count("information") * Count("sugar")) / Total count (W) )
PPMI("information", "sugar") = log2( (5 / 473 * 61) / 11716 )
PPMI("information", "sugar") = log2( 0.00021231 )
PPMI("information", "sugar") ≈ -12.8482
Therefore, the PPMI between "information" and "sugar" is approximately -12.8482.