CNN binary classifier: validation accuracy fluctuation2019 Community Moderator ElectionOverfitting and cross-validationPossible Reason for low Test accuracy and high AUCvalidation/training accuracy and overfittingValidation showing huge fluctuations. What could be the cause?Validation set performance increased, test set performance decreasedConstant validation loss & accuracy, training accuracy fluctuatesissue with early-stopping on f1 score with imbalanced dataValidation accuracy is always close to training accuracyOversampling before Cross-Validation, is it a problem?Stop CNN model at high accuracy and low loss rate?
How would photo IDs work for shapeshifters?
Email Account under attack (really) - anything I can do?
Does the average primeness of natural numbers tend to zero?
What is the meaning of "of trouble" in the following sentence?
LWC and complex parameters
Information to fellow intern about hiring?
Is there any use for defining additional entity types in a SOQL FROM clause?
Is ipsum/ipsa/ipse a third person pronoun, or can it serve other functions?
Manga about a female worker who got dragged into another world together with this high school girl and she was just told she's not needed anymore
Domain expired, GoDaddy holds it and is asking more money
Is domain driven design an anti-SQL pattern?
Creating a loop after a break using Markov Chain in Tikz
Can the Produce Flame cantrip be used to grapple, or as an unarmed strike, in the right circumstances?
Find the positive root of a 4-th degree polynomial equation
Is window.confirm() accessible?
Calculate Levenshtein distance between two strings in Python
Is "plugging out" electronic devices an American expression?
Can I find out the caloric content of bread by dehydrating it?
How to deal with fear of taking dependencies
Can I legally use front facing blue light in the UK?
I see my dog run
New order #4: World
COUNT(*) or MAX(id) - which is faster?
Why do UK politicians seemingly ignore opinion polls on Brexit?
CNN binary classifier: validation accuracy fluctuation
2019 Community Moderator ElectionOverfitting and cross-validationPossible Reason for low Test accuracy and high AUCvalidation/training accuracy and overfittingValidation showing huge fluctuations. What could be the cause?Validation set performance increased, test set performance decreasedConstant validation loss & accuracy, training accuracy fluctuatesissue with early-stopping on f1 score with imbalanced dataValidation accuracy is always close to training accuracyOversampling before Cross-Validation, is it a problem?Stop CNN model at high accuracy and low loss rate?
$begingroup$
For a (binary) classification problem, I made a CNN which works pretty fine. However, to be sure there are no bugs in the data preparation and such, I also trained my network on observations that should NOT be distinguishable, i.e. I would expect a "coin-flip" / 50% accuracy. The total number of observations is several 100k, the data set is balanced and the training - validation split is 80-20.
Seems to me that the model over-fits a bit, but (as expected) does not generalize to the validation set.
The graphical result for a (very long) model training phase looks like this:
Should I be surprised to see (almost) exclusively negative values for my training set accuracy (mean: 0.496, sd: 0.005)? Should it not fluctuate in a more "balanced" way around 0.5?
machine-learning deep-learning cnn training
asked Mar 29 at 10:02
percuissespercuisses
11
$endgroup$
add a comment |
$begingroup$
For a (binary) classification problem, I made a CNN which works pretty fine. However, to be sure there are no bugs in the data preparation and such, I also trained my network on observations that should NOT be distinguishable, i.e. I would expect a "coin-flip" / 50% accuracy. The total number of observations is several 100k, the data set is balanced and the training - validation split is 80-20.
Seems to me that the model over-fits a bit, but (as expected) does not generalize to the validation set.
The graphical result for a (very long) model training phase looks like this:
Should I be surprised to see (almost) exclusively negative values for my training set accuracy (mean: 0.496, sd: 0.005)? Should it not fluctuate in a more "balanced" way around 0.5?
machine-learning deep-learning cnn training
asked Mar 29 at 10:02
percuissespercuisses
11
$endgroup$
add a comment |
$begingroup$
For a (binary) classification problem, I made a CNN which works pretty fine. However, to be sure there are no bugs in the data preparation and such, I also trained my network on observations that should NOT be distinguishable, i.e. I would expect a "coin-flip" / 50% accuracy. The total number of observations is several 100k, the data set is balanced and the training - validation split is 80-20.
Seems to me that the model over-fits a bit, but (as expected) does not generalize to the validation set.
The graphical result for a (very long) model training phase looks like this:
Should I be surprised to see (almost) exclusively negative values for my training set accuracy (mean: 0.496, sd: 0.005)? Should it not fluctuate in a more "balanced" way around 0.5?
machine-learning deep-learning cnn training
asked Mar 29 at 10:02
percuissespercuisses
11
$endgroup$
For a (binary) classification problem, I made a CNN which works pretty fine. However, to be sure there are no bugs in the data preparation and such, I also trained my network on observations that should NOT be distinguishable, i.e. I would expect a "coin-flip" / 50% accuracy. The total number of observations is several 100k, the data set is balanced and the training - validation split is 80-20.
Seems to me that the model over-fits a bit, but (as expected) does not generalize to the validation set.
The graphical result for a (very long) model training phase looks like this:
Should I be surprised to see (almost) exclusively negative values for my training set accuracy (mean: 0.496, sd: 0.005)? Should it not fluctuate in a more "balanced" way around 0.5?
machine-learning deep-learning cnn training
machine-learning deep-learning cnn training
asked Mar 29 at 10:02
percuissespercuisses
11
asked Mar 29 at 10:02
percuissespercuisses
11
asked Mar 29 at 10:02
percuissespercuisses
11
asked Mar 29 at 10:02
percuissespercuisses
11
asked Mar 29 at 10:02
percuissespercuisses
11
11
add a comment |
add a comment |
0
active
oldest
votes
Your Answer
StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
);
);
, "mathjax-editing");
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
);
);
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
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f48206%2fcnn-binary-classifier-validation-accuracy-fluctuation%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
Thanks for contributing an answer to Data Science Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
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
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f48206%2fcnn-binary-classifier-validation-accuracy-fluctuation%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
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
Required, but never shown
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
Required, but never shown
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
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
