Am trying to classify 10000 samples of text into 20 classes. 4 of the classes have just 1 sample each, I tried SMOTE to address this imbalance, but I am unable to generate new samples for classes that have only one record, though I could generate samples for classes with more than 1 sample. Any suggestions?
A good explainer (and a potential answer to your question on why it might not have worked on the undersampled classes) on SMOTE can be found in this answer.
I think this issue can't be solved easily through off-the-shelf data augmentation strategies. One possibility might be to simply duplicate the example, but this would add no new information to your model.
Here are a couple other strategies you could try as well:
An embedding-based augmentation technique (similar theory to SMOTE but works better on text data) that's described in this 2015 paper by William Wang and Diyi Yang.
A step further on #1 using contextualized word embeddings described here in this 2017 paper by Marzieh Fadaee, Arianna Bisazza, and Christof Monz.
Use a synonym replacement library like WordNetAug.
Related
i hope everyone is doing well
I need some help with generative models.
So im working on a project where the main task is to build a binary classification model. In the dataset which contains 300000 sample and 100 feature, there is an imbalance between the 2 classes where majority class is too much bigger than the minory class.
To handle this problem, i'm using VAE (variational autoencoders) to solve this problem.
So i started training the VAE on the minority class and then use the decoder part of the VAE to generate new or fake samples that are similars to the minority class then concatenate this new data with training set in order to have a new balanced training set.
My question is : is there anyway to evalutate generative models like vae, like is there a way to know if the data generated is similar to the real one ??
I have read that there is some metrics to evaluate generated data like inception distance and Frechet inception distance but i saw that they have been only used on image data
I wanna know if i can use them too on my dataset ?
Thanks in advance
I believe your data is not image as you say there are 100 features. What I believe that you can check the similarity between the synthesised features and the original features (the ones belong to minority class), and keep only the ones with certain similarity. Cosine similarity index would be useful for this problem.
That would be also very nice to check a scatter plot of the synthesised features with the original ones to see if they are close to each other. tSNE would be useful at this point.
I have a binary classification problem where I have around 15 features. I have chosen these features using some other model. Now I want to perform Bayesian Logistic on these features. My target classes are highly imbalance(minority class is 0.001%) and I have around 6 million records. I want to build a model which can be trained nighty or weekend using Bayesian logistic.
Currently, I have divided the data into 15 parts and then I train my model on the first part and test on the last part then I am updating my priors using Interpolated method of pymc3 and rerun the model using the 2nd set of data. I am checking the accuracy and other metrics(ROC, f1-score) after each run.
Problems:
My score is not improving.
Am I using the right approch?
This process is taking too much time.
If someone can guide me with the right approach and code snippets it will be very helpful for me.
You can use variational inference. It is faster than sampling and produces almost similar results. pymc3 itself provides methods for VI, you can explore that.
I only know this part of question. If you can elaborate your problem a bit further, maybe.. I can help you.
I am trying to use BERT for a document ranking problem. My task is pretty straightforward. I have to do a similarity ranking for an input document. The only issue here is that I don’t have labels - so it’s more of a qualitative analysis.
I am on my way to try a bunch of document representation techniques - word2vec, para2vec and BERT mainly.
For BERT, i came across Hugging face - Pytorch library. I fine tuned the bert-base-uncased model, with around 150,000 documents. I ran it for 5 epochs, with a batch size of 16 and max seq length 128. However, if I compare the performance of Bert representation vs word2vec representations, for some reason word2vec is performing better for me right now. For BERT, I used the last four layers for getting the representation.
I am not too sure why the fine tuned model didn’t work. I read up this paper, and this other link also that said that BERT performs well when fine tuned for a classification task. However, since I don’t have the labels, I fined tuned it as it's done in the paper - in an unsupervised manner.
Also, my documents vary a lot in their length. So I’m sending them sentence wise right now. In the end I have to average over the word embeddings anyway to get the sentence embedding. Any ideas on a better method? I also read here - that there are different ways of pooling over the word embeddings to get a fixed embedding. Wondering if there is a comparison of which pooling technique works better?
Any help on training BERT better or a better pooling method will be greatly appreciated!
You can check out this blog post:
BERT even has a special [CLS] token whose output embedding is used for classification tasks, but still turns out to be a poor embedding of the input sequence for other tasks. [Reimers & Gurevych, 2019]
Sentence-BERT, presented in [Reimers & Gurevych, 2019] and accompanied by a Python implementation, aims to adapt the BERT architecture by using siamese and triplet network structures to derive semantically meaningful sentence embeddings that can be compared using cosine-similarity
I'm classifying content based on LDA into generic topics such as Music, Technology, Arts, Science
This is the process i'm using,
9 topics -> Music, Technology, Arts, Science etc etc.
9 documents -> Music.txt, Technology.txt, Arts.txt, Science.txt etc etc.
I've filled in each document(.txt file) with about 10,000 lines of content of what i think is "pure" categorical content
I then classify a test document, to see how well the classifier is trained
My Question is,
a.) Is this an efficient way to classify text (using the above steps)?
b.) Where should i be looking for "pure" topical content to fill each of these files? Sources which are not too large (text data > 1GB)
classification is only on "generic" topics such as the above
a) The method you describe sounds fine, but everything will depend on the implementation of labeled LDA that you're using. One of the best implementations I know is the Stanford Topic Modeling Toolbox. It is not actively developed anymore, but it worked great when I used it.
b) You can look for topical content on DBPedia, which has a structured ontology of topics/entities, and links to Wikipedia articles on those topics/entities.
I suggest you to use bag-of-words (bow) for each class you are using. Or vectors where each column is the frequency of important keywords related to the class you want to target.
Regarding the dictionaries you have DBPedia as yves referred or WordNet.
a.)The simplest solution is surely the k-nearest neighbors algorithm (knn). In fact, it will classify new texts with categorical content using an overlap metric.
You could find ressources here: https://github.com/search?utf8=✓&q=knn+text&type=Repositories&ref=searchresults
Dataset issue:
If you are dealing with classifying live user feeds, then I guess no single dataset will suffice your requirement.
Because if new movie X released, it might not catch by your classification dataset as the training dataset is obsoleted for it now.
For classification I guess to stay updated with latest datasets, use twitter training datasets. Develop dynamic algorithm which update the classifier with latest updated tweet datasets. You could select top 15-20 hash tag for each category of your choice to get most relevant dataset for each category.
Classifier:
Most of the classifier uses bag of words model, you can try out various classifiers and see which gives best result. see :
http://www.nltk.org/howto/classify.html
http://scikit-learn.org/stable/supervised_learning.html
I'm new to Topic Models, Classification, etc… now I'm already a while doing a project and read a lot of research papers. My dataset consists out of short messages that are human-labeled. This is what I have come up with so far:
Since my data is short, I read about Latent Dirichlet Allocation (and all it's variants) that is useful to detect latent words in a document.
Based on this I found a Java implementation of JGibbLDA http://jgibblda.sourceforge.net but since my data is labeled, there is an improvement of this called JGibbLabeledLDA https://github.com/myleott/JGibbLabeledLDA
In most of the research papers, I read good reviews about Weka so I messed around with this on my dataset
However, again, my dataset is labeled and therefore I found an extension of Weka called Meka http://sourceforge.net/projects/meka/ that had implementations for Multi-labeled data
Reading about multi-labeled data, I know most used approaches such as one-vs-all and chain classifiers...
Now the reason me being here is because I hope to have an answer to following questions:
Is LDA a good approach for my problem?
Should LDA be used together with a classifier (NB, SVM, Binary Relevance, Logistic Regression, …) or is LDA 'enough' to function as a classifier/estimator for new, unseen data?
How do I need to interpret the output coming from JGibbLDA / JGibbLabeledLDA. How do I get from these files to something which tells me what words/labels are assigned to the WHOLE message (not just to each word)
How can I use Weka/Meka do get to what I want in previous question (in case LDA is not what I'm looking for)
I hope someone, or more than one person, can help me figure out how I need to do this. The general idea of all is not the issue here, I just don't know how to go from literature to practice. Most of the papers don't give enough description of how they perform their experiments OR are too technical for my background about the topics.
Thanks!