I have a binary classification problem where I have a few great features that have the power to predict almost 100% of the test data because the problem is relatively simple.
However, as the nature of the problem requires, I have no luxury to make mistake(let's say) so instead of giving a prediction I am not sure of, I would rather have the output as probability, set a threshold and would be able to say, "if I am less than %95 sure, I will call this "NOT SURE" and act accordingly". Saying "I don't know" rather than making a mistake is better.
So far so good.
For this purpose, I tried Gaussian Bayes Classifier(I have a cont. feature) and Logistic Regression algorithms, which provide me the probability as well as the prediction for the classification.
Coming to my Problem:
GBC has around 99% success rate while Logistic Regression has lower, around 96% success rate. So I naturally would prefer to use GBC.
However, as successful as GBC is, it is also very sure of itself. The odds I get are either 1 or very very close to 1, such as 0.9999997, which makes things tough for me, because in practice GBC does not provide me probabilities now.
Logistic Regression works poor, but at least gives better and more 'sensible' odds.
As nature of my problem, the cost of misclassifying is by the power of 2 so if I misclassify 4 of the products, I lose 2^4 more (it's unit-less but gives an idea anyway).
In the end; I would like to be able to classify with a higher success than Logistic Regression, but also be able to have more probabilities so I can set a threshold and point out the ones I am not sure of.
Any suggestions?
Thanks in advance.
If you have enough data, you can simply retune the probabilities. For example, given the "predicted probability" output of your gaussian classifier, you can go back through (on a held out dataset) and at different prediction values, estimate the probability of the positive class.
Further, you can simply set up an optimization on your holdout set to determine the best threshold(without actually estimating the probability). Since it's one dimensional, you shouldn't even need to do anything fancy for optimization-- test like 500 different thresholds and pick the one which minimizes the costs associated with misclassifications.
Related
I used the three algorithm with the same training and test set. However I'm getting the same exact mean accuracy for all of them. Is there any good explanation on why this is the case for me. I read it may have something to do with the classes being used might be considered easy.
It’s a bit weird that this happens but I have to make assumptions regarding to your question:
The MLP architecture consists of a n inputs and one hidden neuron and m outputs.
You use a linear kernel for the SVM
Your test and training data is linearly separateable.
If 1. is true, your network is equal to logistic regression and your data is separate by a line which leads to the 3. and also to 2. because you won’t need a kernel function to separate the data.
So, the weird part is, that the SVM turns the computation of a decision boundary of your classifier into a convex optimization problem to solve for the optimal boundary. Neither Logistic Regression nor an MLP is able to do this. Hence, your test data must be really easy to separate and must lay with a larger margin to the decision boundary than your training data. This way, it’s not necessary to have a optimal margin between the classes and any boundary which separates the classes without error is sufficient.
They can all give identical performance if your problem is simple enough. There's nothing stopping them from giving you identical results.
When people try to solve the task of semantic segmentation with CNN's they usually use a softmax-crossentropy loss during training (see Fully conv. - Long). But when it comes to comparing the performance of different approaches measures like intersection-over-union are reported.
My question is why don't people train directly on the measure they want to optimize? Seems odd to me to train on some measure during training, but evaluate on another measure for benchmarks.
I can see that the IOU has problems for training samples, where the class is not present (union=0 and intersection=0 => division zero by zero). But when I can ensure that every sample of my ground truth contains all classes, is there another reason for not using this measure?
Checkout this paper where they come up with a way to make the concept of IoU differentiable. I implemented their solution with amazing results!
It is like asking "why for classification we train log loss and not accuracy?". The reason is really simple - you cannot directly train for most of the metrics, because they are not differentiable wrt. to your parameters (or at least do not produce nice error surface). Log loss (softmax crossentropy) is a valid surrogate for accuracy. Now you are completely right that it is plain wrong to train with something that is not a valid surrogate of metric you are interested in, and the linked paper does not do a good job since for at least a few metrics they are considering - we could easily show good surrogate (like for weighted accuracy all you have to do is weight log loss as well).
Here's another way to think about this in a simple manner.
Remember that it is not sufficient to simply evaluate a metric such as accuracy or IoU while solving a relevant image problem. Evaluating the metric must also help the network learn in which direction the weights must be nudged towards, so that a network can learn effectively over iterations and epochs.
Evaluating this direction is what the earlier comments mean that the errors are differentiable. I suppose that there is nothing about the IoU metrics that the network can use to say: "hey, it's not exactly here, but I have to maybe move my bounding box a little to the left!"
Just a trickle of an explanation, but hope it helps..
I always use mean IOU for training a segmentation model. More exactly, -log(MIOU). Plain -MIOU as a loss function will easily trap your optimizer around 0 because of its narrow range (0,1) and thus its steep surface. By taking its log scale, the loss surface becomes slow and good for training.
I have a dataset of 5331 positive and 5331 negative reviews. I want to mark Intensity of each review. Intensity can either be "0" or "1".
Is their any technique that I can manually mark 1000 reviews and train a classifier. If the classifier performs very good (say 90% s-fold validation) then I can fill the remaining review using the classifier's output? Will it be a justified assumption to fill 1/10th of data manually and predict the remaining?
I am new to Machine Learning.
The phrase you are looking for is sentiment analysis and is a well known problem in machine learning society. It is one of the easier tasks of NLP classification, so with great probability you can achieve even more than 90% accuracy. In general scors of 10-CV are a quite reasonable approximation of the real classifier's behaviour, assuming big enough dataset. There are also other (often considered better) techniques, like those based on bootstrap - google for Err^0.632 for an example.
I'm working on a machine learning problem in image processing. I want to get the location of an object in an image by using Histogram of Oriented Gradients (HOG) and a support vector machine (SVM). I've read a couple of articles and tutorials about training the SVM. The setup is pretty standard. I have labeled positive training images and now need to generate a set of negative training samples.
In literature, the approach to generate negative training samples by randomly choosing a position is found very often. I've also seen some approaches where in a successive step to choosing random negative samples, the false-positives of a detection are used as negative training samples once again.
However, I'm wondering if one could not use this approach generally from the start. So one would generate only one false training sample randomly, run the detection and put false-positives in the negative training set again. This seems quite an obvious strategy to me, but I wonder if I'm missing something.
The theory behind this method is laid out in Object Detection with Discriminatively Trained Part Based Models by P. Felzenszwalb, R. Girshick, D. McAllester, D. Ramanan in their PAMI paper. In essence, your starting negative set does not matter, you will always converge to the same classifier if you iteratively add hard samples (with an SVM margin > -1). Starting with a single negative would simply make this convergence slower.
To me it sounds like you want to train the SVM classifier online/incrementally, i.e. updating the classifier with new samples. Such methods are generally only used if new data comes available over time. In your case it seems that you can generate a whole set of negative training samples, so there would be no need to train it incrementally. I'm inclined to say that training the classifier in one run will be better than doing this incrementally (as hinted at by larsmans).
(Again, I'm not an image processing specialist, so take this with a grain of salt.)
I'm wondering if one could not use this approach generally from the start.
You'd need some way to detect the false positives from a classification run. To do so, you need a ground truth, that is, you need a human in the loop. In effect, you'd be doing active learning. If that's what you want to do, you could just as well start with a bunch of hand-labeled negative examples.
Alternatively, you could set this up as a PU learning problem. I have no idea whether that works well with images, but for text classification, it sometimes works.
I created a heuristic (an ANN, but that's not important) to estimate the probabilities of an event (the results of sports games, but that's not important either). Given some inputs, this heuristics tell me what are the probabilities of the event. Something like : Given theses inputs, team B as 65% chances to win.
I have a large set of inputs data for which I now the result (games previously played). Which formula/metric could I use to qualify the accuracy of my estimator.
The problem I see is, if the estimator says the event has a probability of 20% and the event actually do occurs. I have no way to tell if my estimator is right or wrong. Maybe it's wrong and the event was more likely than that. Maybe it's right, the event as about 20% chance to occur and did occur. Maybe it's wrong, the event has really low chances to occurs, say 1 in 1000, but happened to occur this time.
Fortunately I have lots of theses actual test data, so there is probably a way to use them to qualify my heuristic.
anybody got an idea?
There are a number of measurements that you could use to quantify the performance of a binary classifier.
Do you care whether or not your estimator (ANN, e.g.) outputs a calibrated probability or not?
If not, i.e. all that matters is rank ordering, maximizing area under ROC curve (AUROC) is a pretty good summary of the performance of the metric. Others are "KS" statistic, lift. There are many in use, and emphasize different facets of performance.
If you care about calibrated probabilities then the most common metrics are the "cross entropy" (also known as Bernoulli probability/maximum likelihood, the typical measure used in logistic regression) or "Brier score". Brier score is none other than mean squared error comparing continuous predicted probabilites to binary actual outcomes.
Which is the right thing to use depends on the ultimate application of the classifier. For example, your classifier may estimate probability of blowouts really well, but be substandard on close outcomes.
Usually, the true metric that you're trying to optimize is "dollars made". That's often hard to represent mathematically but starting from that is your best shot to coming up with an appropriate and computationally tractable metric.
In a way it depends on the decision function you are using.
In the case of a binary classification task (predicting whether an event occurred or not [ex: win]), a simple implementation is to predict 1 if the probability is greater than 50%, 0 otherwise.
If you have a multiclass problem (predicting which one of K events occurred [ex: win/draw/lose]), you can predict the class with the highest probability.
And the way to evaluate your heuristic is to compute the prediction error by comparing the actual class of each input with the prediction of your heuristic for that instance.
Note that you would usually divide your data into train/test parts to get better (unbiased) estimates of the performance.
Other tools for evaluation exist such as ROC curves, which is a way to depict the performance with regard to true/false postitives.
As you stated, if you predict that an event has a 20% of happening - and 80% not to happen - observing a single isolated event would not tell you how good or poor your estimator was. However, if you had a large sample of events for which you predicted 20% success, but observe that over that sample, 30% succeeded, you could begin to suspect that your estimator is off.
One approach would be to group your events by predicted probability of occurrence, and observe the actual frequency by group, and measure the difference. For instance, depending on how much data you have, group all events where you predict 20% to 25% occurrence, and compute the actual frequency of occurrence by group - and measure the difference for each group. This should give you a good idea of whether your estimator is biased, and possibly for which ranges it's off.