GILMA Lecture Series- Guidelines to identifty whether a girl in Orkut is a figure or not

Guidelines to identifty whether a girl in Orkut is a figure or not

Introduction:

Often, we come across a girl's profile in orkut or any other social networking site and wonder if she is a figure or not. The answer is obvious if she has put her photo but that does not happen in most cases. So, many guys endup sending her a friend request , start chatting and discover very late that she is not. This is a terrible loss of productivity and manhours that could have very well been spent on other girls who are potential figures. It can be easily seen that this can be avoided only if the gilma had been able to identify whether the girl is a figure or not. So this paper presents methods to solve this problem.

Assumptions:

The "relationship status" field of the girl should be "single". There is no point even if the field is left empty.

Preliminary filtering:

Use this method if you have a large number of profiles in hand and you want to filter them into a smaller set so as to better focus your energies on that small set.

1) Reject if the profile picture is that of a baby. Sometimes the girl posts her own baby picture which might be cute, but dont be fooled.

2) Reject if the profile picture has flowers or of actresses.

3) Reject if the profile picture shows only the eyes of girl or any posture which hides her full face.

4) Reject completely if the profile picture is that of Emma Watson (hermione granger in Harry Potter Movie). This rule applies to small sets as well. Even though no logical explanation has been offered for this so far, statistical data reveals that any girl with a picture of emma watson will not be a figure with probability of 0.99.

5) If the profile pic shows a group of people where alteast one of them is not a figure, she is that one. So reject.

6) Reject if the "about me" section of the profile says "Be my friend to know more about me".


Profile classification techniques:

The following methods apply to a smaller set of profiles and takes relatively longer time to classify the profile. This has a score based approach where the readers are asked to compute the score of the profile based on the following rules. The profile is concluded to be that of a figure if the total score exceeds a certain threshold (say 35).

1) For rules 1 & 2, her scrapbook has to be readable by public. +3 score for every scrap in the last 50 scraps in her scrapbook which reads "Hi, didnt u study in X college (or working in Y company)", "Hi, do you want to be my friend" or "Why didnt u come to college ( or office)", "I heard u were sick. Are u better". Note, these scraps would have to be from guys only.

2) For every scrap from a guy which looks like the beginning of a conversation, see who started the conversation (based on timestamp of the scraps). +3 if the guy did it , -3 if the girl did it.

3) +5 for every testimonial written by another guy. -5 if the testimonial says "she is like my sister".

4) Rate of increase of number of friends in her friends list. (Difference between the number of friends on Day1 & Day15 / Number of frnds of Day1 ). Here, the readers may argue that the girl might have been sending out friend requests to a lot of guys. But if that was the case, her profile would have been rejected in preliminary filtering itself. Also, the ratio (rate of increase) is used instead of the actual increase to deal with both the following scenarios

a) Some figs like to see the number of ppl in their frnds list go up since they see it as a matter of prestige. So they will accept most requests. In this case the increase in a given time will be high.
b) Some figs do not accept frnd requests from strangers.These people will have a relatively lower number of people in their list and the rate of increase will also be slow.

Hence, taking the ratio gives equal weight to both the cases.

5) +5 if her friends are hot. Even if she did not turn out to be a figure, her contact might pay good benefits later on.

6) -4 if the girl had been found playing games such "best thing about the profile above", "+2 for girls -2 for boys"...etc. in any of the communities.

Conclusion:
The solutions presented here are merely guidelines based on a number of observations and has been found to work around 85% of the time. Note, there will always be certain exceptions.

Glossary:
figure - hot chick
gilma - Refer to http://www.orkut.co.in/Main#Community.aspx?cmm=38596396

Notes:
More such posts from the gilma team can be found in http://gilma-lectures.blogspot.com/