Please post your solutions to the in-class exercises to this thread. Best, Chris

-- Sep 18 In-Class Exercise Thread

Why I need to search: To know if people around me are mean ;) Ans: Relevant in first 10 : 3 results Precsion @10 = 3/10 = 0.3 Google returns there a 286,000,000 results. I assume 30% are relevant from above question so |Rel| =0.3 * total =85800000

-- Sep 18 In-Class Exercise Thread

Query: mean person behaviour 1. If you think there is some mean person, you want to see what kind of behaviour they have? 2. There is a facebook group with the name mean person behaviour 3. One wants to know mean person behaviour in US Search Engine: Google P@10 = 2/10 = 0.2 |Rel| To estimate the relevance, one can enter the detailed query (add specific words), this way the search engine will return the number of results. This count can be helpful in determining relevance. Another way could be to use the P@10 which shows relevant results. If Google returns 1000 results then relevance can be multiplying P@10 with 1000.

-- Sep 18 In-Class Exercise Thread

'Mean person Behavior' Why - Usual activities, traits and attributes of a mean person. On google, 4 out of top 10 results were relevant to my query. So, |Rel| = 4 Considering k = 10. P@10 = 4/10 = 0.4

-- Sep 18 In-Class Exercise Thread

I would search for "mean person behavior" if i had a bad day if someone is rude to me or someone stole my stuff. I found Rel=6 out of k=10 results as relevant. P@10 = 6/10= 0.6 Total relevant : 1,2,3,5,6,8 irrelevant : 4,7,9,10 So,P@10 = 6/10= 0.6 but there were 286,000,000 results, out of which i only checked the top 10 for relevance.\So, P@10= 0.6 for me.

-- Sep 18 In-Class Exercise Thread

WHY?? - I would write 'mean person behaviour' as a query to know what exactly is meant by mean behaviour and why would someone be mean. Search engine: google.com |Rel| : All the results that would answer my query and intent = 3 k = 10 P@10 = 3/10 = 0.3 Relevance would then be, 0.3 * (the total number of results) In case of Google, 182,000,000 results are returned, so the |Rel| = 0.3*182,000,000

-- Sep 18 In-Class Exercise Thread

1) In order to determine why are people mean 2) Search Engine chosen- Google. Precision- 4/10=0.4 3)Out of 10 results, 4 were relevant and the relevance started decreasing going down the page and on the next pages. Would estimate precision to be ~4/286,000,000 which is very less. :)

-- Sep 18 In-Class Exercise Thread

Why I might enter query "mean person behaviour": to find out how a mean person would react to some situations or to try to understand why some person show some mean behavior toward me. Search engine: Bing P@10 = 3/10 = 0.3 Estimate |Rel|: Bing showing about 29,300,000 Results. The number of relevant pages decrease after each 10 pages, so I would estimate the percent would show 20%. Lots of them shows only topic of behaviour, not mean behaviour. 29,300,000 x 20% = 5860000

-- Sep 18 In-Class Exercise Thread

<nowiki>Why? Reason: One of my friends is being mean to me. I want to know the possible ways to not be bothered by it. Search Engine: Google. Way to estimate relevance: The results which meet the information need. P@10: 0.7 Naive way to estimate |Rel|: Total results returned: 286,000,000 |Rel| = 286,000,000 * 0.7 = 200,200,000</nowiki>

-- Sep 18 In-Class Exercise Thread

          10 --> 0.6

        286,000,000  --- > x

SE: Google Why: - to understand what are the characteristics of a mean person. P@10: i > 1 2 3 4 5 6 7 8 9 10 1|0 > 0 1 1 0 0 1 1 0 1 1 num(rel_results) = 6 P@10 = 6/10 = 0.6 Estimate: let's say total results = 286,000,000 for 10 --> 0.6 then 286,000,000 --- > x x = (0.6 * 286,000,000) / 10 x = 17160000 | Rel | = 17160000