Trjtdtk

What is the evidence that custom checks in Northern Ireland are going to result in violence?

Why did Bronn offer to be Tyrion Lannister's champion in trial by combat?

Can a Knight grant Knighthood to another?

A journey... into the MIND

Protagonist's race is hidden - should I reveal it?

How to keep bees out of canned beverages?

Why aren't road bike wheels tiny?

Providing direct feedback to a product salesperson

Is Bran literally the world's memory?

Why does BitLocker not use RSA?

Compiling and throwing simple dynamic exceptions at runtime for JVM

What kind of equipment or other technology is necessary to photograph sprites (atmospheric phenomenon)

Pointing to problems without suggesting solutions

How do I deal with an erroneously large refund?

Etymology of 見舞い

"Destructive force" carried by a B-52?

Why do people think Winterfell crypts is the safest place for women, children & old people?

A German immigrant ancestor has a "Registration Affidavit of Alien Enemy" on file. What does that mean exactly?

/bin/ls sorts differently than just ls

tabularx column has extra padding at right?

Can this water damage be explained by lack of gutters and grading issues?

What helicopter has the most rotor blades?

Why does my GNOME settings mention "Moto C Plus"?

Converting a text document with special format to Pandas DataFrame

Certainity of a classifier

0

$begingroup$

How to build a classifier that by default will predict that it is for class 1, but if the classifier believes with 80 certainity that it belongs to 0, it will be classed as 0. How to check how certain a classifier is on it's prediction.

python classifier

share|improve this question

edited Apr 4 at 15:43

pcko1

1,726418

asked Apr 4 at 15:32

OmanOman

82

$endgroup$

• 
  
  
  

  
  1
  

  
  
  
  
  
  

  
  $begingroup$
  Why don't you use a classifier that can export probabilities (like a Decision Tree) and make the prediction manually from there? If the probability of class 0 is > 0.8, return 0, else return 1.
  $endgroup$
  – Tasos
  Apr 4 at 15:42
  

  
  
  
  

add a comment | 

0

$begingroup$

How to build a classifier that by default will predict that it is for class 1, but if the classifier believes with 80 certainity that it belongs to 0, it will be classed as 0. How to check how certain a classifier is on it's prediction.

python classifier

share|improve this question

edited Apr 4 at 15:43

pcko1

1,726418

asked Apr 4 at 15:32

OmanOman

82

$endgroup$

• 
  
  
  

  
  1
  

  
  
  
  
  
  

  
  $begingroup$
  Why don't you use a classifier that can export probabilities (like a Decision Tree) and make the prediction manually from there? If the probability of class 0 is > 0.8, return 0, else return 1.
  $endgroup$
  – Tasos
  Apr 4 at 15:42
  

  
  
  
  

add a comment | 

$begingroup$

How to build a classifier that by default will predict that it is for class 1, but if the classifier believes with 80 certainity that it belongs to 0, it will be classed as 0. How to check how certain a classifier is on it's prediction.

python classifier

share|improve this question

edited Apr 4 at 15:43

pcko1

1,726418

asked Apr 4 at 15:32

OmanOman

82

$endgroup$

share|improve this question

edited Apr 4 at 15:43

pcko1

1,726418

asked Apr 4 at 15:32

OmanOman

82

share|improve this question

edited Apr 4 at 15:43

pcko1

1,726418

asked Apr 4 at 15:32

OmanOman

82

edited Apr 4 at 15:43

pcko1

1,726418

edited Apr 4 at 15:43

pcko1

1,726418

asked Apr 4 at 15:32

OmanOman

82

asked Apr 4 at 15:32

OmanOman

82

2 Answers
2

active

oldest

votes

2

$begingroup$

Many classifiers will give the option to get predicted probability. Then you can just put a threshold. Here is how it can be done in with sklearn:

from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification

# Make a dataset
X, y = make_classification(n_samples=1000, n_features=4,
 n_informative=2, n_redundant=0,
 random_state=0, shuffle=False)

clf = RandomForestClassifier(n_estimators=100, max_depth=2,
 random_state=0)
clf.fit(X, y)

# 1 if proba is less than 0.8, otherwise 0
predictions = 1 - (clf.predict_proba(X)[:, 0] > 0.80)

share|improve this answer

edited Apr 4 at 15:55

answered Apr 4 at 15:42

Simon LarssonSimon Larsson

1,100214

$endgroup$

add a comment | 

0

$begingroup$

You can build a neural network with softmax activation on the output layer, to give you values within the range [0,1]. Then you can further post-process those predictions however you like, i.e. using a threshold of 0.8 for binary classification between 0 and 1.

share|improve this answer

answered Apr 4 at 15:41

pcko1pcko1

1,726418

$endgroup$

add a comment | 

Your Answer

StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "557"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);

else
createEditor();

);

function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);



);

draft saved

draft discarded

Sign up or log in

StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);

Sign up using Google

Sign up using Facebook

Sign up using Email and Password

Post as a guest

Name

Email

Required, but never shown

StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f48615%2fcertainity-of-a-classifier%23new-answer', 'question_page');

);

Post as a guest

Name

Email

Required, but never shown

2

$begingroup$

Many classifiers will give the option to get predicted probability. Then you can just put a threshold. Here is how it can be done in with sklearn:

from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification

# Make a dataset
X, y = make_classification(n_samples=1000, n_features=4,
 n_informative=2, n_redundant=0,
 random_state=0, shuffle=False)

clf = RandomForestClassifier(n_estimators=100, max_depth=2,
 random_state=0)
clf.fit(X, y)

# 1 if proba is less than 0.8, otherwise 0
predictions = 1 - (clf.predict_proba(X)[:, 0] > 0.80)

share|improve this answer

edited Apr 4 at 15:55

answered Apr 4 at 15:42

Simon LarssonSimon Larsson

1,100214

$endgroup$

add a comment | 

2

$begingroup$

Many classifiers will give the option to get predicted probability. Then you can just put a threshold. Here is how it can be done in with sklearn:

from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification

# Make a dataset
X, y = make_classification(n_samples=1000, n_features=4,
 n_informative=2, n_redundant=0,
 random_state=0, shuffle=False)

clf = RandomForestClassifier(n_estimators=100, max_depth=2,
 random_state=0)
clf.fit(X, y)

# 1 if proba is less than 0.8, otherwise 0
predictions = 1 - (clf.predict_proba(X)[:, 0] > 0.80)

share|improve this answer

edited Apr 4 at 15:55

answered Apr 4 at 15:42

Simon LarssonSimon Larsson

1,100214

$endgroup$

add a comment | 

$begingroup$

Many classifiers will give the option to get predicted probability. Then you can just put a threshold. Here is how it can be done in with sklearn:

from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification

# Make a dataset
X, y = make_classification(n_samples=1000, n_features=4,
 n_informative=2, n_redundant=0,
 random_state=0, shuffle=False)

clf = RandomForestClassifier(n_estimators=100, max_depth=2,
 random_state=0)
clf.fit(X, y)

# 1 if proba is less than 0.8, otherwise 0
predictions = 1 - (clf.predict_proba(X)[:, 0] > 0.80)

share|improve this answer

edited Apr 4 at 15:55

answered Apr 4 at 15:42

Simon LarssonSimon Larsson

1,100214

$endgroup$

from sklearn.ensemble import RandomForestClassifier
from sklearn.datasets import make_classification

# Make a dataset
X, y = make_classification(n_samples=1000, n_features=4,
 n_informative=2, n_redundant=0,
 random_state=0, shuffle=False)

clf = RandomForestClassifier(n_estimators=100, max_depth=2,
 random_state=0)
clf.fit(X, y)

# 1 if proba is less than 0.8, otherwise 0
predictions = 1 - (clf.predict_proba(X)[:, 0] > 0.80)

share|improve this answer

edited Apr 4 at 15:55

answered Apr 4 at 15:42

Simon LarssonSimon Larsson

1,100214

answered Apr 4 at 15:42

Simon LarssonSimon Larsson

1,100214

answered Apr 4 at 15:42

Simon LarssonSimon Larsson

1,100214

0

$begingroup$

You can build a neural network with softmax activation on the output layer, to give you values within the range [0,1]. Then you can further post-process those predictions however you like, i.e. using a threshold of 0.8 for binary classification between 0 and 1.

share|improve this answer

answered Apr 4 at 15:41

pcko1pcko1

1,726418

$endgroup$

add a comment | 

0

$begingroup$

You can build a neural network with softmax activation on the output layer, to give you values within the range [0,1]. Then you can further post-process those predictions however you like, i.e. using a threshold of 0.8 for binary classification between 0 and 1.

share|improve this answer

answered Apr 4 at 15:41

pcko1pcko1

1,726418

$endgroup$

add a comment | 

$begingroup$

You can build a neural network with softmax activation on the output layer, to give you values within the range [0,1]. Then you can further post-process those predictions however you like, i.e. using a threshold of 0.8 for binary classification between 0 and 1.

share|improve this answer

answered Apr 4 at 15:41

pcko1pcko1

1,726418

$endgroup$

share|improve this answer

answered Apr 4 at 15:41

pcko1pcko1

1,726418

answered Apr 4 at 15:41

pcko1pcko1

1,726418

answered Apr 4 at 15:41

pcko1pcko1

1,726418