explain predictions - value Vs importance of an attribute

Thiru

hi,  I have a question with respect to " Explain predictions " operator. refer the screen shot attached.
shows the output of the operator.
Trying to under stand what difference beween the columns: 

a.    'value' vs ' importance'.  These two values against a row are related?
means  For eg.  Higher the 'value', higher the 'importance'.

b.  How these 'value' and 'importance' are calculated'.  Any web resources/book to understand
the meaning behind them.?  thanks



best regards
thiru

rdesai

Hi Thiru, 

Good observation and nice question! 

Try using Model simulator operator and you will get clear picture how the value and importance columns are related to each other. The slider(in the output of the operator)  takes in the value and show for that particular set of values, what's the importance factor value of each attribute. For E.g The picture you shared, tells you that if the specific set of values(say first 6 rows in your case = one combination of values) is given, then what's the importance of the respected attributes. 

Hope this helps!