I am trying to train a CNN for a multiclass - multilabel classification task (20 classes, each sample can belong to 1+ labels) and the dataset is highly imbalanced. In single-label cases I would use the compute_class_weights function from sklearn to calculate the class weights in order to help the optimizer to account for the minority class. However, for the multilabel case I feel its not working as supposed to, because it considers as number of samples the number of …

Edit: following comment from @anony-mousse, I'm changing the question to search for a general clustering approach that matches this criterion (minimum weight per cluster). I am to use a clustering method on a set of $n$ weighted points: --------------------------------------------- | id | weight | feature_1| feature_2 | ... | --------------------------------------------- | 1 | 4 | 0.2345 | -0.2345 | ... | | 2 | 2 | 0.675 | 0.7433 | ... | | 3 | 15 | -0.45 | 0.123 …

I have a confusion matrix TN= 27 FP=20 FN =11 TP=6 I want to calculate the weighted average for accuracy, sensitivity and specificity. I know the equation but unsure how to do the weighted averages.

I've got a regression problem where a model is required to predict a value in the range [0, 1]. I've tried to look at the distribution of the data and and it seems that there are more examples with a low value label ([0, 0.2]) than higher value labels ([0.2, 1]). When I try to train the model using the MAE metric, the model converges to a state where it has a very low loss, but it seems that the …

For image classification tasks, is there a practical difference between using weighted loss functions vs. using weighted sampling? (I would appreciate theoretical arguments, experience or published papers, anything really.) Some details: By "weighted sampling", I mean attributing different sampling probabilities for each sample in the training set. By "weighted loss functions", I mean weighting error terms differently depending on the sample considered.

I have a model with 2 categorical outputs. The first output layer can predict 2 classes: [0, 1] and the second output layer can predict 3 classes: [0, 1, 2]. How can I apply different class weight dictionaries for each of the outputs? For example, how could I apply the dictionary {0: 1, 1: 10} to the first output, and {0: 5, 1: 1, 2: 10} to the second output? I've tried to use the following class weights dictionary weight_class={'output1': …

My goal is to build a classification model in order to predict if a customer will buy a product or not (binary classification). Since in the last months (let's say 3-4) I know that the advertising of the company is changed a bit, I want to put more emphasis on the newer records. I know that it is possible to specify the sample_weights parameter in most of the classification algorithms, but I don't know how to properly build these weights. …

I have a dataset where each response variable is the number of successes of N Bernoulli trials with N and p (the probability of success) being different for each observation. The goal is to train a model to predict p given the predictors. However observations with a small N will have a higher variance and higher N. Consider the following scenario to illustrate better: Assume coins with different pictures on them have a different bias and that the bias is …

I'm currently studying Boosting techniques in Machine Learning and I happened to understand that in Algorithms like Adaboost, each of the training samples is given a weight depending on whether it was misclassified or not by the previous model in sequential boosting. Although I intuitively understand that by weighting examples, we are letting the model pay more attention to examples that were previously misclassified, I do not understand "how" the weights are taken into account by a machine learning algorithm. …

In a classification problem, is it suitable to assign sample weights based on their positive class probability? For example, if I am building a binary classification problem where one of the independent features has three possible values a – 2% of the samples, probability for positive class = 90% b – 8% of the samples, probability for positive class = 40% c – 90% of the samples, probability for positive class = 5% Can I assign the samples weights based …

I am trying to understand how to build custom layers in Keras and I went through a couple examples: here and here. The syntax is, of course, similar, but in non of the cases it is addressed why weights are declared the way they are. For instance, how do I constrain weights to be binary? Thanks in advance!

I know there is a similar Qn at Unbalanced multiclass data with XGBoost. But I don't understand the reply provided by @Esmailian. What is the actual formula to obtain 1, 0.333 and 0.167? For example, if we have three imbalanced classes with ratios class A = 10% class B = 30% class C = 60% Their weights would be (dividing the smallest class by others) class A = 1.000 class B = 0.333 class C = 0.167 Will I obtain …

Across a few different projects, I have hit a problem where I have two (or more) models: General-Purpose Model: A model which is based on a large amount of data not specifically relevant to my current classifier label goal, but which predict other labels using similar features. Cold-Start Model: A model trained on data specifically related to my current label/task, which initially starts with zero observations and goes up from there. So then, my question: what is an appropriate way …

I am performing a logistic regression in a standard supervised framework (Data Set X, target y). The dataset X is composed of a handfull of categorical variables (that I one-hot encode), thus it contains a lot of redundant rows (1000s unique rows over millions of initial rows). Having a lot of redundant rows I was tempted to agregate them, weight them by their count in the fit and get approximately the same result. However I was surprised to get variation …

there is a option sample_weight in fit(X[, y, sample_weight]) function (OneClassSVM, sklearn library). If I use the option sample_weight , I might give some weight to some point(that are likely to be more normal points), right? Otherwise, what does mean the sample_weight? link: https://scikit-learn.org/stable/modules/generated/sklearn.svm.OneClassSVM.html#sklearn.svm.OneClassSVM.score_samples

I have a dataset (around 45,000 screenshots) of UI elements (UI trees containing element types and bounding boxes) and associated screenshots: The dataset is highly imbalanced with the button element being highly overrepresented: When training on my local machine on a tiny subset of the data (900 screenshots for training, 100 for testing) and 10 epochs, my results aren't bad: I trained the model on Azure ML with 25,000 screenshots for 13 epochs (which took about 3 days) and my …

I know that there is a possibility in Keras with the class_weights parameter dictionary at fitting, but I couldn't find any example. Would somebody so kind to provide one? By the way, in this case the appropriate praxis is simply to weight up the minority class proportionally to its underrepresentation?

I have a dataset with a few strongly imbalanced classes, eg. the smallest class is about 54 times smaller than the largest. Therefore, data augmentation in order to equalize the size of classes seems like a bad idea to me (in the example above each image would have to be augmented 54 times on average). So I thought that I could do less augmentation of minority classes and then use class weights in the loss function. Is this approach better …

I realized I need to use the package survey to be able to include sample weights in my regression analysis. Initially, I wanted to use a negative binomial regression on each one of my outcomes as count data is more often than not overdispersed, so I tried using svyglm.nb. However, for one of the outcomes which has small values, svyglm.nb makes my program crash, so I think there might be some convergence issue. I thought using a Poisson regression might …

Similiar to class imbalance there is always something I would call "learnability imbalance" in multi-class classification. What I mean by that: Even when the classes are evenly distributed in the dataset some classes will be classified more easily by the model than others. An example would be a CNN model that classifies dog, cat and car. Dog and cat will most likely have a lower true positive rate than car because cats and dogs look more similiar to each other. …