we have a news website where we have to match news to a particular user.
We have to use for the matching only the user textual information, like for example the interests of the user or a brief description about them.
I was thinking to threat both the user textual information and the news text as document and find document similarity.
In this way, I hope, that if in my profile I wrote sentences like: I loved the speach of the president in Chicago last year, and a news talks about: Trump is going to speak in Illinois I can have a match (the example is purely casual).
I tried, first, to embed my documents using TF-IDF and then I tried a kmeans to see if there was something that makes sense, but I don't like to much the results.
I think the problem derives from the poor embedding that TF-IDF gives me.
Thus I was thinking of using BERT embedding to retrieve the embedding of my documents and then use cosine similarity to check similarity of two document (a document about the user profile and a news).
Is this an approach that could make sense? Bert can be used to retrieve the embedding of sentences, but there is a way to embed an entire document?
What would you advice me?
Thank you
BERT is trained on pairs of sentences, therefore it is unlikely to generalize for much longer texts. Also, BERT requires quadratic memory with the length of the text, using too long texts might result in memory issues. In most implementations, it does not accept sequences longer than 512 subwords.
Making pre-trained Transformers work efficiently for long texts is an active research area, you can have a look at a paper called DocBERT to have an idea what people are trying. But it will take some time until there is a nicely packaged working solution.
There are also other methods for document embedding, for instance Gensim implements doc2vec. However, I would still stick with TF-IDF.
TF-IDF is typically very sensitive to data pre-processing. You certainly need to remove stopwords, in many languages it also pays off to do lemmatization. Given the specific domain of your texts, you can also try expanding the standard list of stop words by words that appear frequently in news stories. You can get further improvements by detecting and keeping together named entities.
Related
I am working on a Named Entity Recognition (NER) project in which I got a large amount of text in the sense that it is too much to read or skim read. Therefore, I want to create an overview of what is mentioned by extracting named entities (places, names, times, maybe topics) and create an index of kind (entity, list of pages/lines where it is mentioned). I have worked through Standford's NLP lecture, (parts of) Eisenstein's Introduction to NLP book found some literature and systems for English texts. As my corpus is in German, I would like to ask how I can approach this problem. Also, this is my first NLP project, so I would not know if I could solve this challenge even if texts were in English.
As a first step
are there German NER systems out there which I could use?
The further roadmap of my project is:
How can I avoid mapping misspellings or rare names to a NUL/UNK token? This is relevant because there are also some historic passages that use words no longer in use or that follow old orthography. I think the relevant terms are tokenisation or stemming.
I thought about fine-tuning or transfer learning the base NER model to a corpus of historic texts to improve NER.
A major challenge is that there is no annotated dataset for my corpus available and I could only manually annotate a tiny fraction of it. So I would be happy for hints on German annotated datasets which I could incorporate into my project.
Thank you in advance for your inputs and fruitful discussions.
Most good NLP toolkits can perform NER in German:
Stanford NLP
Spacy
probably NLTK and OpenNLP as well
What is crucial to understand is that using NER software like the above means using a pretrained model, i.e. a model which has been previously trained on some standard corpus with standard annotated entities.
Btw you can usually find the original annotated dataset by looking at the documentation. There's one NER corpus here.
This is convenient and might suit your goal, but sometimes it doesn't collect exactly every that you would like it to collect, especially if your corpus is from a very specific domain. If you need more specific NER, you must train your own model and this requires obtaining some annotated data (i.e. manually annotating or paying somebody to do it).
Even in this case, a NER model is statistical and it will unavoidably make some mistakes, don't expect perfect results.
About misspellings or rare names: a NER model doesn't care (or not too much) about the actual entity, because it's not primarily based on the words in the entity. It's based on indications in the surrounding text, for example in the sentence "It was announced by Mr XYZ that the event would take place in July", the NER model should find 'Mr XYZ' as a person due to "announced by" and 'July' as a date because of "take place in". However if the language used in the corpus is very different from the training data used for the model, the performance could be very bad.
Does anyone know how of an accurate tool or method that can be used to compute word embeddings or find similarity among domain-specific words? I'm working on an NLP project that involves computing cosine similarity between technical terms, such as "address" and "socket", but pre-trained models like word2vec aren't giving useful embeddings or accurate cosine similarities because they aren't specific to technical terms. Since the more general-nontechnical meanings of "address" and "socket" aren't similar to one another, these pretrained models aren't giving them sufficiently high similarity scores for the purposes of my project. Would appreciate any advice people would be able to offer. Thank you!
With sufficient data from your specific domain, you can train your own word2vec model - whose resulting word-vectors, being only influenced by your domain data, will be far more reflective of the in-domain meanings.
Similarly, if you have a mixture of data where you have hints that some word uses are for different senses of a polysemous word, you could try preprocessing your text, using those hints, replacing the ambiguous tokens (like say 'address') with a larger number of distinct tokens (like 'address*networking', 'address*delivery', etc). Even with a lot of error in such a process, its results might be sufficient for a specific purpose.
For example, maybe you'd assume all docs of a certain type – like articles from a particular publication – always mean 'address*networking' when they write 'address'. That crude replacement, on just some subset of docs sufficient to collect enough varied examples of 'address*networking' usage, might leave you with a good-enough word-vector for 'address*networking'.
(More generally, deciding which word sense of multiple candidates is meant by a particular word is called "word sense disambiguation", and it might be possible to use other preexisting code for performing that to help preprocess texts - replacing ambiguous tokens with more-speciific stand-ins – before performing word2vec training.)
Even without such assistive pre-processing, there've been a number of research attempts to extend word2vec to better model words with multiple contrasting meanings. Googling for [word2vec polysemy] or [polysemous embeddings] should turn up a bunch of examples.
But I don't know any of those techniques that have become widely-used, or that are explicitly supported by major word2vec libraries, so I can't specifically recommend or show working code for any. I don't know a standard best-practice or off-the-shelf solution – you'd have to treat adopting those ideas from research papers as an R&D project, performing a lot of your own implementation/evaluation to see if any help with your goals.
How to find list of Out Of Vocabulary (OOV) words from my domain spectific pdf while using FastText model? I need to fine tune FastText with my domain specific words.
A FastText model will already be able to generate vectors for OOV words.
So there's not necessarily any need to either list the specifically OOV words in your PDF, nor 'fine tune' as FastText model.
You just ask it for vectors, it gives them back. The vectors for full in-vocabulary words, that were trained from relevant training material, will likely be best, while vectors synthesized for OOV words from word-fragments (character n-grams) shared with training material will just be rough guesses - better than nothing, but not great.
(To train a good word-vector requires many varied examples of a word's use, interleaved with similarly good examples of its many 'peer' words – and traditionally, in one unified, balanced training session.)
If you think you need to do more, you should expand your questin with more details about why you think that's necessary, and what existing precedents (in docs/tutorials/papers) you're trying to match.
I've not seen a well-documented way to casually fine-tune, or incrementally expand the known-vocabulary of, an existing FastText model. There would be a lot of expert tradeoffs required, and in many cases simply training a new model with sufficient data is likely to be a safer approach.
Anyone seeking such fine-tuning should have a clear idea of:
what their incremental data might be able to add to an existing model
what process/code will they be using, and why that process/code might be expected to give meaningful results with their specific starting model & new data
how the results of any such process can be evaluated to ensure the extra fine-tuning steps are beneficial compared to alternatives
So I am doing a project on document similarity and right now my features are only the embeddings from Doc2Vec. Since that is not showing any good results, after hyperparameter optimization and word embedding before the doc embedding... What other features can I add, so as to get better results?
My dataset is 150 documents, 500-700 words each, with 10 topics(labels), each document having one topic. Documents are labeled on a document level, and that labeling is currently used only for evaluation purposes.
Edit: The following is answer to gojomo's questions and elaborating on my comment on his answer:
The evaluation of the model is done on the training set. I am comparing if the label is the same as the most similar document from the model. For this I am first getting the document vector using the model's method 'infer_vector' and then 'most_similar' to get the most similar document. The current results I am getting are 40-50% of accuracy. A satisfactory score would be of at least 65% and upwards.
Due to the purpose of this research and it's further use case, I am unable to get a larger dataset, that is why I was recommended by a professor, as this is a university project, to add some additional features to the document embeddings of Doc2Vec. As I had no idea what he ment, I am asking the community of stackoverflow.
The end goal of the model is to do clusterization of the documents, again the labels for now being used only for evaluation purposes.
If I don't get good results with this model, I will try out the simpler ones mentioned by #Adnan S #gojomo such as TF-IDF, Word Mover's Distance, Bag of words, just presumed I would get better results using Doc2Vec.
You should try creating TD-IDF with 2 and 3 grams to generate a vector representation for each document. You will have to train the vocabulary on all the 150 documents. Once you have the TF-IDF vector for each document, you can use cosine similarity between any two of them.
Here is a blog article with more details and doc page for sklearn.
How are you evaluating the results as not good, and how will you know when your results are adequate/good?
Note that just 150 docs of 400-700 words each is a tiny, tiny dataset: typical datasets used published Doc2Vec results include tens-of-thousands to millions of documents, of hundreds to thousands of words each.
It will be hard for any of the Word2Vec/Doc2Vec/etc-style algorithms to do much with so little data. (The gensim Doc2Vec implementation includes a similar toy dataset, of 300 docs of 200-300 words each, as part of its unit-testing framework, and to eke out even vaguely useful results, it must up the number of training epochs, and shrink the vector size, significantly.)
So if intending to use Doc2Vec-like algorithms, your top priority should be finding more training data. Even if, in the end, only ~150 docs are significant, collecting more documents that use similar domain language can help improve the model.
It's unclear what you mean when you say there are 10 topics, and 1 topic per document. Are those human-assigned categories, and are those included as part of the training texts or tags passed to the Doc2Vec algorithm? (It might be reasonable to include it, depending on what your end-goals and document-similarity evaluations consist of.)
Are these topics the same as the labelling you also mention, and are you ultimately trying to predict the topics, or just using the topics as a check of the similarity-results?
As #adnan-s suggests in the other answer, it may also be worth trying more-simple count-based 'bag of words' document representations, including potentially on word n-grams or even character n-grams, or TF-IDF weighted.
If you have adequate word-vectors, as trained from your data or from other compatible sources, the "Word Mover's Distance" measure can be another interesting way to compute pairwise similarities. (However, it may be too expensive to calculate between many-hundred-word texts - working much faster on shorter texts.)
As others have already suggested your training set of 150 documents probably isn't big enough to create good representations. You could, however, try to use a pre-trained model and infer the vectors of your documents.
Here is a link where you can download a (1.4GB) DBOW model trained on English Wikipedia pages, working with 300-dimensional document vectors. I obtained the link from jhlau/doc2vec GitHub repository. After downloading the model you can use it as follows:
from gensim.models import Doc2Vec
# load the downloaded model
model_path = "enwiki_dbow/doc2vec.bin"
model = Doc2Vec.load(model_path)
# infer vector for your document
doc_vector = model.infer_vector(doc_words)
Where doc_words is a list of words in your document.
This, however, may not work for you in case your documents are very specific. But you can still give it a try.
I am given a task of classifying a given news text data into one of the following 5 categories - Business, Sports, Entertainment, Tech and Politics
About the data I am using:
Consists of text data labeled as one of the 5 types of news statement (Bcc news data)
I am currently using NLP with nltk module to calculate the frequency distribution of every word in the training data with respect to each category(except the stopwords).
Then I classify the new data by calculating the sum of weights of all the words with respect to each of those 5 categories. The class with the most weight is returned as the output.
Heres the actual code.
This algorithm does predict new data accurately but I am interested to know about some other simple algorithms that I can implement to achieve better results. I have used Naive Bayes algorithm to classify data into two classes (spam or not spam etc) and would like to know how to implement it for multiclass classification if it is a feasible solution.
Thank you.
In classification, and especially in text classification, choosing the right machine learning algorithm often comes after selecting the right features. Features are domain dependent, require knowledge about the data, but good quality leads to better systems quicker than tuning or selecting algorithms and parameters.
In your case you can either go to word embeddings as already said, but you can also design your own custom features that you think will help in discriminating classes (whatever the number of classes is). For instance, how do you think a spam e-mail is often presented ? A lot of mistakes, syntaxic inversion, bad traduction, punctuation, slang words... A lot of possibilities ! Try to think about your case with sport, business, news etc.
You should try some new ways of creating/combining features and then choose the best algorithm. Also, have a look at other weighting methods than term frequencies, like tf-idf.
Since your dealing with words I would propose word embedding, that gives more insights into relationship/meaning of words W.R.T your dataset, thus much better classifications.
If you are looking for other implementations of classification you check my sample codes here , these models from scikit-learn can easily handle multiclasses, take a look here at documentation of scikit-learn.
If you want a framework around these classification that is easy to use you can check out my rasa-nlu, it uses spacy_sklearn model, sample implementation code is here. All you have to do is to prepare the dataset in a given format and just train the model.
if you want more intelligence then you can check out my keras implementation here, it uses CNN for text classification.
Hope this helps.