Linear regression prediction don't match with the model

PlatyQ

Hello

This time I have a question about linear regression operator.

There is my process: I want predict a value (AverageW) with 3 parameters (Layers, WFS, TS) and observe the model choose by the operator.

<?xml version="1.0" encoding="UTF-8"?><process version="8.2.000">
 <context>
 <input/>
 <output/>
 <macros/>
 </context>
 <operator activated="true" class="process" compatibility="8.2.000" expanded="true" name="Process">
 <parameter key="random_seed" value="-1"/>
 <process expanded="true">
 <operator activated="true" class="retrieve" compatibility="8.2.000" expanded="true" height="68" name="Retrieve TS50+80" width="90" x="45" y="34">
 <parameter key="repository_entry" value="../data/TS50+80"/>
 </operator>
 <operator activated="true" class="select_attributes" compatibility="8.2.000" expanded="true" height="82" name="Select Attributes AW" width="90" x="179" y="34">
 <parameter key="attribute_filter_type" value="subset"/>
 <parameter key="attributes" value="Layers|TS|WFS|AverageW"/>
 </operator>
 <operator activated="true" class="set_role" compatibility="8.2.000" expanded="true" height="82" name="Set Role AW" width="90" x="313" y="34">
 <parameter key="attribute_name" value="AverageW"/>
 <parameter key="target_role" value="label"/>
 <list key="set_additional_roles"/>
 </operator>
 <operator activated="true" class="split_data" compatibility="8.2.000" expanded="true" height="103" name="Split Data" width="90" x="447" y="34">
 <enumeration key="partitions">
 <parameter key="ratio" value="0.75"/>
 <parameter key="ratio" value="0.25"/>
 </enumeration>
 <parameter key="sampling_type" value="stratified sampling"/>
 <parameter key="use_local_random_seed" value="true"/>
 <parameter key="local_random_seed" value="1"/>
 </operator>
 <operator activated="true" class="linear_regression" compatibility="8.2.000" expanded="true" height="103" name="Linear Regression" width="90" x="648" y="34"/>
 <operator activated="true" class="apply_model" compatibility="8.2.000" expanded="true" height="82" name="Apply Model" width="90" x="648" y="238">
 <list key="application_parameters"/>
 </operator>
 <connect from_op="Retrieve TS50+80" from_port="output" to_op="Select Attributes AW" to_port="example set input"/>
 <connect from_op="Select Attributes AW" from_port="example set output" to_op="Set Role AW" to_port="example set input"/>
 <connect from_op="Set Role AW" from_port="example set output" to_op="Split Data" to_port="example set"/>
 <connect from_op="Split Data" from_port="partition 1" to_op="Linear Regression" to_port="training set"/>
 <connect from_op="Split Data" from_port="partition 2" to_op="Apply Model" to_port="unlabelled data"/>
 <connect from_op="Linear Regression" from_port="model" to_op="Apply Model" to_port="model"/>
 <connect from_op="Apply Model" from_port="labelled data" to_port="result 1"/>
 <connect from_op="Apply Model" from_port="model" to_port="result 2"/>
 <portSpacing port="source_input 1" spacing="0"/>
 <portSpacing port="sink_result 1" spacing="0"/>
 <portSpacing port="sink_result 2" spacing="0"/>
 <portSpacing port="sink_result 3" spacing="0"/>
 </process>
 </operator>
</process>

The problem is that I obtain a prediction that don't match with the result that I obtain with coefficients give by the model.

For example to coefficients:

coef_Layer = 0.150;

coef_TS = -0.045;

coef_WFS = 1.150;

intercept = 2.488 ;

And example Layer=2 ; TS= 50; WFS= 3
I compute Layer*coef_Layer + coef_TS*TS + coef_WFS*WFS + intercept = 3,988

but the model predict 3.968 to this example!

It is not a big difference but I need to understand if I forget a parameter "epsilon" or other.

I hope somebody can help me because I don't find a answer in documentation (And it is not the first time I have question about documentation)

My data are in the table below if there are problems with csv file:

(I remove not used column so select attributes is not useful)

Spoiler

Layers	WFS	TS	AverageW
1	3	50,0	3,0
1	4	50,0	4,0
1	5	50,0	4,1
1	6	50	7,2
2	3	50,0	3,9
2	4	50,0	4,9
2	5	50,0	5,3
2	6	50,0	7,5
3	3	50	4,3
3	4	50,0	5,4
3	5	50,0	5,8
3	6	50,0	7,6
5	3	50,0	4,5
5	4	50	6,3
5	5	50,0	6,9
5	6	50,0	10,8
10	3	50,0	5,0
10	4	50,0	6,7
10	5	50	8,1
20	3	50,0	5,5
20	4	50,0	7,3
20	5	50,0	9,1
1	3	80,0	2,4
1	4	80,0	3,7
1	5	80,0	3,7
1	6	80,0	4,7
2	3	80,0	3,1
2	4	80,0	4,1
2	5	80,0	4,1
2	6	80,0	5,8
3	3	80,0	3,3
3	4	80,0	4,0
3	5	80,0	4,5
3	6	80,0	6,9
5	3	80,0	3,7
5	4	80,0	4,6
5	5	80,0	5,1
5	6	80,0	6,9
10	3	80,0	3,8
10	4	80,0	5,2
10	5	80,0	6,4

Thank you in advance

 

lionelderkrikor

Hi @PlatyQ,

 

In reality the displayed coefficients are rounded : 

 

So if you perform the calculus with the precise numbers, you obtain 3.968 : 

 

 

Regards,

 

Lionel

PlatyQ

@lionelderkrikor

thank, I never running the mouse over the table enough time to see precise numbers.

Thank you!