Laughing
Well-known member
Cheers Devon, Bear, interesting thoughts. My interest comes form machine leaning only and I tend to feed all data and make no presumptive conclusions. You are tight Bear that you need a lot of data, 10's of thousands of rows for training a bayes net to any decent level so you are right with the data we have available it is abacus ratherr than deep blue.
What a bayes net does really well though is when you can rule out a factor, the human brain continues to use that bit of data, orphaned data essentially, to make decisions. A bayes net doesn't do that. A really simple example would be someone brought up as catholic who later renounces that religion, will still have orphaned assumptions based on the upbringing, for example, sex outside of marriage is wrong. They are likely to maintain that viewpoint despite the fact that the underlying reason is removed from their decision making.
What a bayes net does really well though is when you can rule out a factor, the human brain continues to use that bit of data, orphaned data essentially, to make decisions. A bayes net doesn't do that. A really simple example would be someone brought up as catholic who later renounces that religion, will still have orphaned assumptions based on the upbringing, for example, sex outside of marriage is wrong. They are likely to maintain that viewpoint despite the fact that the underlying reason is removed from their decision making.