Bayesian model of predicting lesbian sub-gender type: PHT# LES - 同女之舞
s*n
1 楼
What proposed here is a Bayesian model of predicting whether a particular
person is P, H or T.
Let Type denote a discreet variable of predictions P, H or T.
Let s denote an observed variable of appearance or attribute.
P(Type) is the probability of T, P, H based on common sense and statistics
without knowing anything about the particular person.
P(s,Type) is the probability of seeing characteristics s (long hair, shorts
, red Jeep, worrying about marrying a man, etc) given the Type.
Bayesian inference tells us:
P(Type= PHT,s ) = P(Type)*P(s,Type)
We want to find out whether our hypothesis, say, an ID is T or not, we just
have to write down the prior distribution and the likelihood.
P(Type) is prior distribution of the type. In north america, we can assume
a skewed distribution that there are much more T than P+H. We can also have
a Prior over geographical distribution. For example, There are more Ts in
Southern California than New York, etc.
How do we get the prior distribution? Very easy. We just take some data from
this BBS Id and that will give us a distribution.
P(s,Type) is the likelihood. It defines, given a particular Type, such as T,
how do we describe the corresponding attribute. The attribute could
include measurable appearance and personality characteristics. For example,
if know a person is T, we can describe THE PROBABILITY of what clothes she
would wear, hairstyle she would have, whether she will carry heavy bags for
her lady friends, etc. On the other hand, if a person is T, how likely he
would prefer dressing in shorts and drive a red Jeep, etc.
Once we have both prior distribution and the likelihood, we can apply
standard machine learning algorithm to find the posterior distribution P(
Type, s).
person is P, H or T.
Let Type denote a discreet variable of predictions P, H or T.
Let s denote an observed variable of appearance or attribute.
P(Type) is the probability of T, P, H based on common sense and statistics
without knowing anything about the particular person.
P(s,Type) is the probability of seeing characteristics s (long hair, shorts
, red Jeep, worrying about marrying a man, etc) given the Type.
Bayesian inference tells us:
P(Type= PHT,s ) = P(Type)*P(s,Type)
We want to find out whether our hypothesis, say, an ID is T or not, we just
have to write down the prior distribution and the likelihood.
P(Type) is prior distribution of the type. In north america, we can assume
a skewed distribution that there are much more T than P+H. We can also have
a Prior over geographical distribution. For example, There are more Ts in
Southern California than New York, etc.
How do we get the prior distribution? Very easy. We just take some data from
this BBS Id and that will give us a distribution.
P(s,Type) is the likelihood. It defines, given a particular Type, such as T,
how do we describe the corresponding attribute. The attribute could
include measurable appearance and personality characteristics. For example,
if know a person is T, we can describe THE PROBABILITY of what clothes she
would wear, hairstyle she would have, whether she will carry heavy bags for
her lady friends, etc. On the other hand, if a person is T, how likely he
would prefer dressing in shorts and drive a red Jeep, etc.
Once we have both prior distribution and the likelihood, we can apply
standard machine learning algorithm to find the posterior distribution P(
Type, s).