I'm looking for an explanation for the following issue. I'm fitting an XGB model to a dataset of 50k rows, 26 features.
I'm running a gridsearch with varying max_depths + n_estimators and the model is performing much better with deeper trees (with a depth of 14 im getting c.87 accuracy, c.83 precision, when I reduce the depth to 5 I have a reduced performance of .85 accuracy and .81 precision). On the validation set, the performance is the same for both models (.81 accuracy, .78 precision).
So at surface level it seems like the deeper model is performing the same or better, but when I plot learning curves for the two models, the deeper model looks like its overfitting. The image below shows the learning curves for the two models with the deeper trees at the top and the shallower trees at the bottom.
How can this be explained?
One very simple reason may be the capacity. The first model, the deeper one, has much more capacity. Thus, due to the capacity it can learn the useful things, leading to the good performance, and after a while it will start to learn things that lead to the overfitting. This usually happens to models with big capacity.
Your second model doesn't have enough capacity to overfit, which may also lead to the fact that it does not have enough capacity to learn useful things.
You can also see on the plots that the bigger model starts to overfit around epoch 10, but it already managed to obtain a lower loss than the second model.
Related
I am reading the a deep learning with python book.
After reading chapter 4, Fighting Overfitting, I have two questions.
Why might increasing the number of epochs cause overfitting?
I know increasing increasing the number of epochs will involve more attempts at gradient descent, will this cause overfitting?
During the process of fighting overfitting, will the accuracy be reduced ?
I'm not sure which book you are reading, so some background information may help before I answer the questions specifically.
Firstly, increasing the number of epochs won't necessarily cause overfitting, but it certainly can do. If the learning rate and model parameters are small, it may take many epochs to cause measurable overfitting. That said, it is common for more training to do so.
To keep the question in perspective, it's important to remember that we most commonly use neural networks to build models we can use for prediction (e.g. predicting whether an image contains a particular object or what the value of a variable will be in the next time step).
We build the model by iteratively adjusting weights and biases so that the network can act as a function to translate between input data and predicted outputs. We turn to such models for a number of reasons, often because we just don't know what the function is/should be or the function is too complex to develop analytically. In order for the network to be able to model such complex functions, it must be capable of being highly-complex itself. Whilst this complexity is powerful, it is dangerous! The model can become so complex that it can effectively remember the training data very precisely but then fail to act as an effective, general function that works for data outside of the training set. I.e. it can overfit.
You can think of it as being a bit like someone (the model) who learns to bake by only baking fruit cake (training data) over and over again – soon they'll be able to bake an excellent fruit cake without using a recipe (training), but they probably won't be able to bake a sponge cake (unseen data) very well.
Back to neural networks! Because the risk of overfitting is high with a neural network there are many tools and tricks available to the deep learning engineer to prevent overfitting, such as the use of dropout. These tools and tricks are collectively known as 'regularisation'.
This is why we use development and training strategies involving test datasets – we pretend that the test data is unseen and monitor it during training. You can see an example of this in the plot below (image credit). After about 50 epochs the test error begins to increase as the model has started to 'memorise the training set', despite the training error remaining at its minimum value (often training error will continue to improve).
So, to answer your questions:
Allowing the model to continue training (i.e. more epochs) increases the risk of the weights and biases being tuned to such an extent that the model performs poorly on unseen (or test/validation) data. The model is now just 'memorising the training set'.
Continued epochs may well increase training accuracy, but this doesn't necessarily mean the model's predictions from new data will be accurate – often it actually gets worse. To prevent this, we use a test data set and monitor the test accuracy during training. This allows us to make a more informed decision on whether the model is becoming more accurate for unseen data.
We can use a technique called early stopping, whereby we stop training the model once test accuracy has stopped improving after a small number of epochs. Early stopping can be thought of as another regularisation technique.
More attempts of decent(large number of epochs) can take you very close to the global minima of the loss function ideally, Now since we don't know anything about the test data, fitting the model so precisely to predict the class labels of the train data may cause the model to lose it generalization capabilities(error over unseen data). In a way, no doubt we want to learn the input-output relationship from the train data, but we must not forget that the end goal is for the model to perform well over the unseen data. So, it is a good idea to stay close but not very close to the global minima.
But still, we can ask what if I reach the global minima, what can be the problem with that, why would it cause the model to perform badly on unseen data?
The answer to this can be that in order to reach the global minima we would be trying to fit the maximum amount of train data, this will result in a very complex model(since it is less probable to have a simpler spatial distribution of the selected number of train data that is fortunately available with us). But what we can assume is that a large amount of unseen data(say for facial recognition) will have a simpler spatial distribution and will need a simpler Model for better classification(I mean the entire world of unseen data, will definitely have a pattern that we can't observe just because we have an access small fraction of it in the form of training data)
If you incrementally observe points from a distribution(say 50,100,500, 1000 ...), we will definitely find the structure of the data complex until we have observed a sufficiently large number of points (max: the entire distribution), but once we have observed enough points we can expect to observe the simpler pattern present in the data that can be easily classified.
In short, a small fraction of train data should have a complex structure as compared to the entire dataset. And overfitting to the train data may cause our model to perform worse on the test data.
One analogous example to emphasize the above phenomenon from day to day life is as follows:-
Say we meet N number of people till date in our lifetime, while meeting them we naturally learn from them(we become what we are surrounded with). Now if we are heavily influenced by each individual and try to tune to the behaviour of all the people very closely, we develop a personality that closely resembles the people we have met but on the other hand we start judging every individual who is unlike me -> unlike the people we have already met. Becoming judgemental takes a toll on our capability to tune in with new groups since we trained very hard to minimize the differences with the people we have already met(the training data). This according to me is an excellent example of overfitting and loss in genralazition capabilities.
I am using the random forest.My test accuracy is 70% on the other hand train accuracy is 34% ? what to do ? How can I solve this problem.
Test accuracy should not be higher than train since the model is optimized for the latter. Ways in which this behavior might happen:
you did not use the same source dataset for test. You should do a proper train/test split in which both of them have the same underlying distribution. Most likely you provided a completely different (and more agreeable) dataset for test
an unreasonably high degree of regularization was applied. Even so there would need to be some element of "test data distribution is not the same as that of train" for the observed behavior to occur.
The other answers are correct in most cases. But I'd like to offer another perspective. There are specific training regimes that could cause the training data to be harder for the model to learn - for instance, adversarial training or adding Gaussian noise to the training examples. In these cases, the benign test accuracy could be higher than train accuracy, because benign examples are easier to evaluate. This isn't always a problem, however!
If this applies to you, and the gap between train and test accuracies is larger than you'd like (~30%, as in your question, is a pretty big gap), then this indicates that your model is underfitting to the harder patterns, so you'll need to increase the expressibility of your model. In the case of random forests, this might mean training the trees to a higher depth.
First you should check the data that is used for training. I think there is some problem with the data, the data may not be properly pre-processed.
Also, in this case, you should try more epochs. Plot the learning curve to analyze when the model is going to converge.
You should check the following:
Both training and validation accuracy scores should increase and loss should decrease.
If there is something wrong in step 1 after any particular epoch, then train your model until that epoch only, because your model is over-fitting after that.
Based on data that our business department supplied to us, I used the sklearn decision tree algorithm to determine the ROC_AUC for a binary classification problem.
The data consists of 450 rows and there are 30 features in the data.
I used 10 times StratifiedKFold repetition/split of training and test data. As a result, I got the following ROC_AUC values:
0.624
0.594
0.522
0.623
0.585
0.656
0.629
0.719
0.589
0.589
0.592
As I am new in machine learning, I am unsure whether such a variation in the ROC_AUC values can be expected (with minimum values of 0.522 and maximum values of 0.719).
My questions are:
Is such a big variation to be expected?
Could it be reduced with more data (=rows)?
Will the ROC_AUC variance get smaller, if the ROC_AUC gets better ("closer to 1")?
Well, you do k-fold splits to actually evaluate how well your model generalizes.
Therefore, from your current results I would assume the following:
This is a difficult problem, the AUCs are usually low.
0.71 is an outlier, you were just lucky there (probably).
Important questions that will help us help you:
What is the proportion of the binary classes? Are they balanced?
What are the features? Are they all continuous? If categorical, are they ordinal or nominal?
Why Decision Tree? Have you tried other methods? Logistic Regression for instance is a good start before you move on to more advanced ML methods.
You should run more iterations, instead of k fold use the ShuffleSplit function and run at least 100 iterations, compute the Average AUC with 95% Confidence Intervals. That will give you a better idea of how well the models perform.
Hope this helps!
Is such a big variation to be expected?
This is a textbook case of high variance.
Depending on the difficulty of your problem, 405 training samples may not be enough for it to generalize properly, and the random forest may be too powerful.
Try adding some regularization, by limiting the number of splits that the trees are allowed to make. This should reduce the variance in your model, though you might expect a potentially lower average performance.
Could it be reduced with more data (=rows)?
Yes, adding data is the other popular way of lowering the variance of your model. If you're familiar with deep learning, you'll know that deep models usually need LOTS of samples to learn properly. That's because they are very powerful models with an intrinsically high variance, and therefore a lot of data is needed for them to generalize.
Will the ROC_AUC variance get smaller, if the ROC_AUC gets better ("closer to 1")?
Variance will decrease with regularization and adding data, it has no relation to the actual performance "number" that you get.
Cheers
Generally speaking, is it possible to tell if training a given neural network of depth X on Y training examples for Z epochs is likely to overfit? Or can overfitting only be detected for sure by looking at loss and accuracy graphs of training vs test set?
Concretely I have ~250,000 examples, each of which is a flat image 200x200px. The model is a CNN with about 5 convolution + pooling layers, followed by 2 dense layers with 1024 units each. The model classifies 12 different classes. I've been training it for about 35 hours with ~90% accuracy on training set and ~80% test set.
Generally speaking, is it possible to tell if training a given neural network of depth X on Y training examples for Z epochs is likely to overfit?
Generally speaking, no. Fitting deep learning models is still an almost exclusively empirical art, and the theory behind it is still (very) poor. And although by gaining more and more experience one is more likely to tell beforehand if a model is prone to overfit, the confidence will generally be not high (extreme cases excluded), and the only reliable judge will be the experiment.
Elaborating a little further: if you take the Keras MNIST CNN example and remove the intermediate dense layer(s) (the previous version of the script used to include 2x200 dense layers instead of 1x128 now), thus keeping only conv/pooling layers and the final softmax one, you will end up with ~ 98.8% test accuracy after only 20 epochs, but I am unaware of anyone that could reliably predict this beforehand...
Or can overfitting only be detected for sure by looking at loss and accuracy graphs of training vs test set?
Exactly, this is the only safe way. The telltale signature of overfitting is the divergence of the learning curves (training error still decreasing, while validation or test error heading up). But even if we have diagnosed overfitting, the cause might not be always clear-cut (see a relevant question and answer of mine here).
~90% accuracy on training set and ~80% test set
Again very generally speaking and only in principle, this does not sound bad for a problem with 12 classes. You already seem to know that, if you worry for possible overfitting, it is the curves rather than the values themselves (or the training time) that you have to monitor.
On the more general topic of the poor theory behind deep learning models as related to the subject of model intepretability, you might find this answer of mine useful...
I have a dataset of approx. 4800 rows with 22 attributes, all numerical, describing mostly the geometry of rock / minerals, and 3 different classes.
I tried out a cross validation with k-nn Model inside it, with k= 7 and Numerical Measure -> Camberra Distance as parameters set..and I got a performance of 82.53% and 0.673 kappa. Is that result representative for the dataset? I mean 82% is quite ok..
Before doing this, I evaluated the best subset of attributes with a decision table, I got out 6 different attributes for that.
the problem is, you still don't learn much from that kind of models, like instance-based k-nn. Can I get any more insight from knn? I don't know how to visualize the clusters in that high dimensional space in Rapidminer, is that somehow possible?
I tried decision tree on the data, but I got too much branches (300 or so) and it looked all too messy, the problem is, all numerical attributes have about the same mean and distribution, therefore its hard to get a distinct subset of meaningful attributes...
ideally, the staff wants to "Learn" something about the data, but my impression is, that you cannot learn much meaningful of that data, all that works best is "Blackbox" Learning models like Neural Nets, SVM, and those other instance-based models...
how should I proceed?
Welcome to the world of machine learning! This sounds like a classic real-world case: we want to make firm conclusions, but the data rows don't cooperate. :-)
Your goal is vague: "learn something"? I'm taking this to mean that you're investigating, hoping to find quantitative discriminations among the three classes.
First of all, I highly recommend Principal Component Analysis (PCA): find out whether you can eliminate some of these attributes by automated matrix operations, rather than a hand-built decision table. I expect that the messy branches are due to unfortunate choice of factors; decision trees work very hard at over-fitting. :-)
How clean are the separations of the data sets? Since you already used Knn, I'm hopeful that you have dense clusters with gaps. If so, perhaps a spectral clustering would help; these methods are good at classifying data based on gaps between the clusters, even if the cluster shapes aren't spherical. Interpretation depends on having someone on staff who can read eigenvectors, to interpret what the values mean.
Try a multi-class SVM. Start with 3 classes, but increase if necessary until your 3 expected classes appear. (Sometimes you get one tiny outlier class, and then two major ones get combined.) The resulting kernel functions and the placement of the gaps can teach you something about your data.
Try the Naive Bayes family, especially if you observe that the features come from a Gaussian or Bernoulli distribution.
As a holistic approach, try a neural net, but use something to visualize the neurons and weights. Letting the human visual cortex play with relationships can help extract subtle relationships.