Trjtdtk

Error message with tabularx

To say I met a person for the first time

Mac Pro install disk keeps ejecting itself

Binary Numbers Magic Trick

What does the "ep" capability mean?

Size of electromagnet needed to replicate Earth's magnetic field

Examples of subgroups where it's nontrivial to show closure under multiplication?

How to get a plain text file version of a CP/M .BAS (M-BASIC) program?

Does Gita support doctrine of eternal cycle of birth and death for evil people?

Repelling Blast: Must targets always be pushed back?

Will tsunami waves travel forever if there was no land?

French for 'It must be my imagination'?

What makes accurate emulation of old systems a difficult task?

Unexpected email from Yorkshire Bank

Who is the Umpire in this picture?

Is there really no use for MD5 anymore?

Is the claim "Employers won't employ people with no 'social media presence'" realistic?

Why does academia still use scientific journals and not peer-reviewed government funded alternatives?

Seemingly unused edef prior to an ifx mysteriously affects the outcome of the ifx. Why?

How can the Zone of Truth spell be defeated without the caster knowing?

Why do games have consumables?

What route did the Hindenburg take when traveling from Germany to the U.S.?

Why other Westeros houses don't use wildfire?

Was there a shared-world project before "Thieves World"?

Number of iterations for minimal cost complexity prunning?

Why is the number of samples smaller than the number of values in my decision tree?Minimum number of trees for Random Forest classifierSpace complexity of classification algorithmsHow to control number of branches for each split in sklearn DT classifier?Decision Trees How to Calculate Entropy Gain for Continuous Values for split point?How to reduce the number of rules in decision tree with Support and ConfidenceHow important is it for each row of data to have the same number of features?Decision Tree Classifier For Minimizing Arbitrary Cost FunctionBuilding a minimal encoding of nominal labels from numerical featuresHow to determine number of leaves in decision tree analysis?

1

$begingroup$

I've been fiddling with weka's J48 decision tree implementation (C4.5). My goal is to implement cost complexity prunning using weakest link cut method. Basically my algorithm iteratively prunes the tree creating trees with fewer nodes from the previous ones ($ T_0 > T_1 > T_2 >...>T_N$).

My goal now is how do I find that $N$ from $T_N$ meaning how many times should I prune the tree. The algorithm seems to prune pretty aggressively, few iterations too much and it will be completely prunned.

For now I'm using the stop condition regarding the complexity factor and I stop the prunning when $alpha gt 0.001$ (from $R(t) = R(t) + alpha|T|$) . This works pretty well for the two datasets I've tested so far but I wouldn't bet my money on it.

decision-trees

share|improve this question

asked Apr 7 at 7:37

oren revengeoren revenge

1062

$endgroup$

add a comment | 

1

$begingroup$

I've been fiddling with weka's J48 decision tree implementation (C4.5). My goal is to implement cost complexity prunning using weakest link cut method. Basically my algorithm iteratively prunes the tree creating trees with fewer nodes from the previous ones ($ T_0 > T_1 > T_2 >...>T_N$).

My goal now is how do I find that $N$ from $T_N$ meaning how many times should I prune the tree. The algorithm seems to prune pretty aggressively, few iterations too much and it will be completely prunned.

For now I'm using the stop condition regarding the complexity factor and I stop the prunning when $alpha gt 0.001$ (from $R(t) = R(t) + alpha|T|$) . This works pretty well for the two datasets I've tested so far but I wouldn't bet my money on it.

decision-trees

share|improve this question

asked Apr 7 at 7:37

oren revengeoren revenge

1062

$endgroup$

add a comment | 

$begingroup$

I've been fiddling with weka's J48 decision tree implementation (C4.5). My goal is to implement cost complexity prunning using weakest link cut method. Basically my algorithm iteratively prunes the tree creating trees with fewer nodes from the previous ones ($ T_0 > T_1 > T_2 >...>T_N$).

My goal now is how do I find that $N$ from $T_N$ meaning how many times should I prune the tree. The algorithm seems to prune pretty aggressively, few iterations too much and it will be completely prunned.

For now I'm using the stop condition regarding the complexity factor and I stop the prunning when $alpha gt 0.001$ (from $R(t) = R(t) + alpha|T|$) . This works pretty well for the two datasets I've tested so far but I wouldn't bet my money on it.

decision-trees

share|improve this question

asked Apr 7 at 7:37

oren revengeoren revenge

1062

$endgroup$

share|improve this question

asked Apr 7 at 7:37

oren revengeoren revenge

1062

share|improve this question

asked Apr 7 at 7:37

oren revengeoren revenge

1062

asked Apr 7 at 7:37

oren revengeoren revenge

1062

asked Apr 7 at 7:37

oren revengeoren revenge

1062