Discussion of Assignment 1 Part II. The output shows the results of running spam_probability three times on each of the six test messages. As can be seen, the output for the first, second, third, and fifth messages is fairly consistent - the estimated probability of the message being spam is about 0.37 for the first message, about 0.60 for the second message, and about 1 for the third and fifth mesages. However, the spam probability output for the fourth and sixth messages is much more variable. This is because the number of generated messages that match the values of the observed variable for these messages is quite small, as you can see by adding a "print(match_count)" statement just before the end of spam_probability. Sometimes there are no matches at all, in which case the estimated spam probability is "NaN" - the result of dividing 0 by 0.