topic How to represent the Y response of the decision tree as a variable? in Discussions

How to represent the Y response of the decision tree as a variable?

lwx228 — Sat, 18 Aug 2018 14:21:52 GMT

dear all

I want to do multiple data table files with the same code for the decision tree operation, and each file has three different Y response decision trees.

Set the Y response of each file in the first 3 columns of the table, and the number of X factor column starting from column 4 is different for each file.

The specific practice of making three decision trees for each file is: the first time is carried out with the Y response of column 1;The second time was done with the Y response of column 2;The third time is done with the Y response of column 3.

The X factor used each time starts with column 5 and ends with column 2.The second X factor to the second X factor.

How to code this decision tree code so that the same code can perform operations in different files.

In the annex, Boston Housing.jmp adds 2 columns to the Y response so that 3 columns of the Y response are available and moved to the first 3 columns of the table.

It's complicated, thank you very much!

Re: How to represent the Y response of the decision tree as a variable?

lwx228 — Sat, 18 Aug 2018 14:26:25 GMT

Upload a data table adjusted by Boston housing.jmp, and I won't upload the code.Thanks!

Re: How to represent the Y response of the decision tree as a variable?

lwx228 — Sat, 18 Aug 2018 14:46:03 GMT

Note: only one data table file is uploaded here. It is assumed that there are several other different data table files in the same path, and the names of Y response and X factor of different data table files are different.In the previous setting, the respective Y responses were all in the first 3 columns, and the number of X factor columns was different.

Re: How to represent the Y response of the decision tree as a variable?

txnelson — Sat, 18 Aug 2018 15:21:34 GMT

Here is a simple script that will run the 3 Decision Trees once the data table is in the form of the first 3 columns being the Y responses, and the remaining columns being the X factors. I do not follow your

"The X factor used each time starts with column 5 and ends with column 2.The second X factor to the second X factor.

How to code this decision tree code so that the same code can perform operations in different files.

In the annex, Boston Housing.jmp adds 2 columns to the Y response so that 3 columns of the Y response are available and moved to the first 3 columns of the table.".

If what I have provided is not what you are seeking, please provide more specifics

Names Default To Here( 1 );
dt = Current Data Table();

// Get all of the columm names
colNamesList = dt << get column names;

// Get the Y Factor columns
yFactorList = As List( colNamesList[1 :: 3] );

// Get the X Factor columns
xFactorList = As List( colNamesList[4 :: N Items( colNamesList )] );

// Run the Partitions


Partition(
	Y( Eval( yFactorList ) ),
	X( Eval( xFactorList ) ),
	Split Best( 3 )
);

Re: How to represent the Y response of the decision tree as a variable?

lwx228 — Sun, 19 Aug 2018 12:46:14 GMT

Thank you. Your answer has solved my problem.

I made a mistake in the previous statement, which should be: X factor goes from column 5 to the penultimate column 2.