LADTree Calculation and How to read the tree

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

LADTree Calculation and How to read the tree

lululala
hi im new in data mining. im still working in my project using weka data
mining software using LADTree classifier. but i cant interprete the tree and
i didnt understand the calculation this classifier used. although ive
already read the journal Multiclass Alternating Decision Trees but i still
confuse how the can get the value in  every node. help please


can someone explain this algorithm
<https://weka.8497.n7.nabble.com/file/t6944/Screenshot_2.png>

=== Classifier model (full training set) ===

weka.classifiers.trees.LADTree:

: 0,0,0
|  (1)CCI < -143: -1,-0.813,1.812
|  |  (5)R3 < 324.5: -0.469,2.901,-2.432
|  |  (5)R3 >= 324.5: -0.484,-0.492,0.975
|  (1)CCI >= -143: -0.814,1.814,-1
|  |  (2)CCI < 199: -0.377,0.88,-0.503
|  |  |  (6)DI 14+ < 44: -0.284,0.752,-0.468
|  |  |  (6)DI 14+ >= 44: 2.893,-2.44,-0.453
|  |  (2)CCI >= 199: 2.769,-2.271,-0.497
|  |  |  (4)MA < 306: -2.439,2.893,-0.455
|  |  |  (4)MA >= 306: 1.104,-0.643,-0.461
|  |  (3)DI 14- < 6.5: 1.59,-1.049,-0.541
|  |  |  (7)S2 < 340.5: 2.89,-2.443,-0.447
|  |  |  (7)S2 >= 340.5: 0.014,0.446,-0.46
|  |  (3)DI 14- >= 6.5: -0.372,0.854,-0.481
|  |  |  (8)CCI < 199: -0.446,0.891,-0.445
|  |  |  (8)CCI >= 199: 0.858,-0.403,-0.455
|  |  (9)R3 < 447.5: -0.453,0.9,-0.447
|  |  (9)R3 >= 447.5: 0.728,-0.255,-0.473
|  |  |  (10)R2 < 459.5: 2.835,-2.392,-0.444
|  |  |  (10)R2 >= 459.5: -0.419,0.867,-0.448
Legend: SELL, HOLD, BUY
#Tree size (total): 31
#Tree size (number of predictor nodes): 21
#Leaves (number of predictor nodes): 13
#Expanded nodes: 100
#Processed examples: 12988
#Ratio e/n: 129.88

and this tree
<https://weka.8497.n7.nabble.com/file/t6944/Screenshot_1.png>




--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
Reply | Threaded
Open this post in threaded view
|

Re: LADTree Calculation and How to read the tree

Eibe Frank-2
Administrator
The classifier produced by LADTree is a so-called alternating decision tree. It has decision nodes, just like a regular decision tree, shown as ellipses in the graphical output, and so-called prediction nodes, shown as rectangles in the graphical output. At a decision node, you must follow the branch whose condition is satisfied by the particular instance you are trying to classify. At a prediction, in contrast, you must always follow all branches extending from the prediction node. Each prediction node also has a vector of numeric values, one for each class value (three in your case). Making a prediction is easy: simply sum up all vectors from all prediction nodes that you end up visiting for the instance you are trying to classify. Then, predict the class value that corresponds to the largest value in the resulting vector.

Cheers,
Eibe

On Sun, Sep 8, 2019 at 6:54 PM lululala <[hidden email]> wrote:
hi im new in data mining. im still working in my project using weka data
mining software using LADTree classifier. but i cant interprete the tree and
i didnt understand the calculation this classifier used. although ive
already read the journal Multiclass Alternating Decision Trees but i still
confuse how the can get the value in  every node. help please


can someone explain this algorithm
<https://weka.8497.n7.nabble.com/file/t6944/Screenshot_2.png>

=== Classifier model (full training set) ===

weka.classifiers.trees.LADTree:

: 0,0,0
|  (1)CCI < -143: -1,-0.813,1.812
|  |  (5)R3 < 324.5: -0.469,2.901,-2.432
|  |  (5)R3 >= 324.5: -0.484,-0.492,0.975
|  (1)CCI >= -143: -0.814,1.814,-1
|  |  (2)CCI < 199: -0.377,0.88,-0.503
|  |  |  (6)DI 14+ < 44: -0.284,0.752,-0.468
|  |  |  (6)DI 14+ >= 44: 2.893,-2.44,-0.453
|  |  (2)CCI >= 199: 2.769,-2.271,-0.497
|  |  |  (4)MA < 306: -2.439,2.893,-0.455
|  |  |  (4)MA >= 306: 1.104,-0.643,-0.461
|  |  (3)DI 14- < 6.5: 1.59,-1.049,-0.541
|  |  |  (7)S2 < 340.5: 2.89,-2.443,-0.447
|  |  |  (7)S2 >= 340.5: 0.014,0.446,-0.46
|  |  (3)DI 14- >= 6.5: -0.372,0.854,-0.481
|  |  |  (8)CCI < 199: -0.446,0.891,-0.445
|  |  |  (8)CCI >= 199: 0.858,-0.403,-0.455
|  |  (9)R3 < 447.5: -0.453,0.9,-0.447
|  |  (9)R3 >= 447.5: 0.728,-0.255,-0.473
|  |  |  (10)R2 < 459.5: 2.835,-2.392,-0.444
|  |  |  (10)R2 >= 459.5: -0.419,0.867,-0.448
Legend: SELL, HOLD, BUY
#Tree size (total): 31
#Tree size (number of predictor nodes): 21
#Leaves (number of predictor nodes): 13
#Expanded nodes: 100
#Processed examples: 12988
#Ratio e/n: 129.88

and this tree
<https://weka.8497.n7.nabble.com/file/t6944/Screenshot_1.png>




--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

_______________________________________________
Wekalist mailing list -- [hidden email]
Send posts to: To unsubscribe send an email to [hidden email]
To subscribe, unsubscribe, etc., visit %(web_page_url)slistinfo%(cgiext)s/%(_internal_name)s
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html