Trjtdtk

How to assert on pagereference where the endpoint of pagereference is predefined

When and why did journal article titles become descriptive, rather than creatively allusive?

What are the spoon bit of a spoon and fork bit of a fork called?

Meaning of "individuandum"

How to creep the reader out with what seems like a normal person?

Has any spacecraft ever had the ability to directly communicate with civilian air traffic control?

Is it always OK to ask for a copy of the lecturer's slides?

How did Captain America use this power?

Can PCs use nonmagical armor and weapons looted from monsters?

How do you center multiple equations that have multiple steps?

Proof that when f'(x) < f(x), f(x) =0

If an enemy is just below a 10-foot-high ceiling, are they in melee range of a creature on the ground?

Catholic vs Protestant Support for Nazism in Germany

Attending a conference where my ex-supervisor and his collaborator are present, should I attend?

Airbnb - host wants to reduce rooms, can we get refund?

Is thermodynamics only applicable to systems in equilibrium?

What word means "to make something obsolete"?

Visa for volunteering in England

I caught several of my students plagiarizing. Could it be my fault as a teacher?

If Melisandre foresaw another character closing blue eyes, why did she follow Stannis?

Accidentally deleted the "/usr/share" folder

Unidentified items in bicycle tube repair kit

Integral of (1/100-x)

How could a planet have most of its water in the atmosphere?

Isolation Forest Implementation

validation_curve differs from cross_val_score?Isolation forest results every value -1Isolation Forest Feature ImportanceWhy is my Keras model not learning image segmentation?Is there any implementation of Extended Isolation Forest algorithm in R/Python?SciKit-Learn Decision Tree OverfittingModelling a startup's funding journey with Brownian Motionpredicting loan default based on transaction historyIsolation Forest Prediction

0

$begingroup$

I implemented Isolation Forest(Liu et al. 2008), but when testing with different datasets using 5 fold cross validation, in kddcup99 dataset(around 4.5 lakh rows), I am getting maximum of 70-72% accuracy, and one kddcup99 available with 2 lakh dataset, I am getting maximum 85-87% accuracy but the author's have given AUC 1.

For satellite I am getting avg of 65% varying subssmpling size from (1 to 96) and above 70% when subsampling size is more than 500, but they have given 71%(1 to 100).

For breast cancer dataset I am getting maximum of 75% accuracy but they have given AUC 0.97.

I have implemented exactly same but don't know why this. Can someone please help me. I can share my accuracies as well.

Thank you.

python

share|improve this question

asked Apr 10 at 10:43

ShivanyaShivanya

166

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Are you working with the same tech stack as the paper ?
  $endgroup$
  – Shamit Verma
  Apr 10 at 12:27
  

  
  
  
  


• 
  
  
  

  
  1
  

  
  
  
  
  
  

  
  $begingroup$
  why do you compare accuracy and AUC? what are your AUC results?
  $endgroup$
  – oW_♦
  Apr 10 at 15:07
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Yes I have been Working on the same work. How to find the actual AUC using formulas. Is always AUC would be more?
  $endgroup$
  – Shivanya
  Apr 10 at 18:01
  
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Is it not (TP+TN) / total
  $endgroup$
  – Shivanya
  Apr 10 at 18:05
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Are you using the same feature preprocessing?
  $endgroup$
  – jonnor
  Apr 10 at 21:57
  

  
  
  
  

 | 
show 6 more comments

0

$begingroup$

I implemented Isolation Forest(Liu et al. 2008), but when testing with different datasets using 5 fold cross validation, in kddcup99 dataset(around 4.5 lakh rows), I am getting maximum of 70-72% accuracy, and one kddcup99 available with 2 lakh dataset, I am getting maximum 85-87% accuracy but the author's have given AUC 1.

For satellite I am getting avg of 65% varying subssmpling size from (1 to 96) and above 70% when subsampling size is more than 500, but they have given 71%(1 to 100).

For breast cancer dataset I am getting maximum of 75% accuracy but they have given AUC 0.97.

I have implemented exactly same but don't know why this. Can someone please help me. I can share my accuracies as well.

Thank you.

python

share|improve this question

asked Apr 10 at 10:43

ShivanyaShivanya

166

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Are you working with the same tech stack as the paper ?
  $endgroup$
  – Shamit Verma
  Apr 10 at 12:27
  

  
  
  
  


• 
  
  
  

  
  1
  

  
  
  
  
  
  

  
  $begingroup$
  why do you compare accuracy and AUC? what are your AUC results?
  $endgroup$
  – oW_♦
  Apr 10 at 15:07
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Yes I have been Working on the same work. How to find the actual AUC using formulas. Is always AUC would be more?
  $endgroup$
  – Shivanya
  Apr 10 at 18:01
  
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Is it not (TP+TN) / total
  $endgroup$
  – Shivanya
  Apr 10 at 18:05
  

  
  
  
  


• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Are you using the same feature preprocessing?
  $endgroup$
  – jonnor
  Apr 10 at 21:57
  

  
  
  
  

 | 
show 6 more comments

$begingroup$

I implemented Isolation Forest(Liu et al. 2008), but when testing with different datasets using 5 fold cross validation, in kddcup99 dataset(around 4.5 lakh rows), I am getting maximum of 70-72% accuracy, and one kddcup99 available with 2 lakh dataset, I am getting maximum 85-87% accuracy but the author's have given AUC 1.

For satellite I am getting avg of 65% varying subssmpling size from (1 to 96) and above 70% when subsampling size is more than 500, but they have given 71%(1 to 100).

For breast cancer dataset I am getting maximum of 75% accuracy but they have given AUC 0.97.

I have implemented exactly same but don't know why this. Can someone please help me. I can share my accuracies as well.

Thank you.

python

share|improve this question

asked Apr 10 at 10:43

ShivanyaShivanya

166

$endgroup$

share|improve this question

asked Apr 10 at 10:43

ShivanyaShivanya

166

share|improve this question

asked Apr 10 at 10:43

ShivanyaShivanya

166

asked Apr 10 at 10:43

ShivanyaShivanya

166

asked Apr 10 at 10:43

ShivanyaShivanya

166