Skip to main content
12 votes
Accepted

Why should the initialization of weights and bias be chosen around 0?

Assuming fairly reasonable data normalization, the expectation of the weights should be zero or close to it. It might be reasonable, then, to set all of the initial weights to zero because a positive ...
Eumenedies's user avatar
11 votes
Accepted

Lemmatization Vs Stemming

I would say that lemmatization is generally the preferred way of reducing related words to a common base. This Quora question is a good resource on the subject: Is it advisable to choose ...
Simon Larsson's user avatar
5 votes
Accepted

NLP - How to perform semantic analysis?

With your three labels: positive, neutral or negative - it seems you are talking more about sentiment analysis. This answer the question: what are the emotions of the person who wrote this piece of ...
n1k31t4's user avatar
  • 14.9k
5 votes

How do I assess which sentiment classifier is best for my project?

A couple of important points: Sentiment analysis is not an exact science. Two people, reading the same text in different contexts will come to different conclusions about sentiment, especially on ...
Neil Slater's user avatar
4 votes

Why should the initialization of weights and bias be chosen around 0?

If you set it as 0, they will all have the same error so backprop will make them all equal; therefore, you should have random initialisation. Why around 0? I think this post may answer it well:
Landmaster's user avatar
4 votes

How do I assess which sentiment classifier is best for my project?

You can classify a few of the tweets yourself, and compare afterwards which of the two algorithmic results is closer to your classification. Without more information we cannot tell what these ...
knb's user avatar
  • 602
4 votes

What algorithm is used to extract keywords from unstructured texts?

The OP asks two different questions: (1) how to extract key words and (2) how to assign keywords a sentiment class (pos/neg/neu). I will address the keyword identification piece in this answer as many ...
Brandon Loudermilk's user avatar
4 votes
Accepted

Interpretation of the loss function for word2vec

Not sure what the video said, but $T$ should not be the vocabulary size, but the training corpus size (number of all words). For example, if your training corpus is ...
user12075's user avatar
  • 2,284
4 votes
Accepted

Is there any named entity reconginition algorithm trained for the french language?

Yes, there is a french model free and ready to use via the spaCy package! Here are the small amd medium sized models, that should be ready to go. Here is the basic summary of the dataset, shown at ...
n1k31t4's user avatar
  • 14.9k
3 votes
Accepted

Why would you use word embeddings to find similar words?

It depends on how similarity is defined. If similarity is defined as human-defined semantics, then a synset (i.e., synonym set) is most appropriate. If similarity is defined as frequent co-occurrence, ...
Brian Spiering's user avatar
3 votes
Accepted

What dataset was Stanford NER trained on?

The original paper mentions two corpora: CoNLL 2003 (apparently here now) and the "CMU Seminar Announcements Task". However according to the page linked in the question the actual NER was trained on a ...
Erwan's user avatar
  • 25.5k
3 votes
Accepted

TF-IDF for Topic Modeling

Formally the problem of topic modelling is a clustering problem: given a collection of text documents, group together the documents which are topically similar. So technically it can indeed be done ...
Erwan's user avatar
  • 25.5k
2 votes

Name Tagger in Stanford NLP

Name parsing does not appear to built-in to Stanford CoreNLP. . One option is writing a series of Regular Expression using Stanford TokensRegex to parse and label name tokens. Another option is ...
Brian Spiering's user avatar
2 votes

Extracting Part of Speech (Source and Destinations) using text mining/NLP?

Named Entity Recognition is technique which can be used here. Location is one of the 3 most studied classes (with Person and Organization). Stanford NLP has an open source Java implementation that is ...
Abhishek Verma's user avatar
2 votes
Accepted

Correcting ALL CAPS for human and algorithmic consumption

If sensitivity to case is breaking your models you have two options: Train or find a new model that's case-insensitive. This is probably the easiest thing to do. The Stanford parser has one. Train a ...
polm23's user avatar
  • 343
2 votes

StanfordTokenizer will be deprecated in version 3.2.5 Warning

@imoutidi, I also encountered the same deprecation warning. After digging around a bit, it looks like the new/replacement package can be imported with the following: ...
groxli's user avatar
  • 21
2 votes
Accepted

What is the more natural parsing, the one that leads to the preferred reading of the sentence

Can anyone explain to me, what is more natural in English and why ? This is a classic example of PP-attachment ambiguity (PP = prepositional phrase). For a full overview of the problem and some ...
JordiCarrera's user avatar
2 votes

How to stay up to date in NLP and use the best approaches?

If you want to be up to date with the new advancements, a good way is skimming through the accepted papers of the major NLP conferences, namely ACL, EMNLP, and the regional EACL, NAACL, AACL. If you ...
noe's user avatar
  • 27k
2 votes
Accepted

How to evaluate triple extraction in NLP?

The standard evaluation method works for this kind of task: measure precision, recall and F1-score on a manually annotated sample. In general one can find which evaluation measure is standard for a ...
Erwan's user avatar
  • 25.5k
1 vote

How to extract the positions of employee from raw text

Have a look at this to see how a Knowledge Base can be built. I would say the best way is to build your own Knowledge Base based on your corpus if you have enough data. The idea (simplified) for ...
Kasra Manshaei's user avatar
1 vote

ImportError: cannot import name 'StanfordCoreNLPParser'

I do not know of anything called StanfordCoreNLPParser. The stanfordcorenlp package has ...
Brian Spiering's user avatar
1 vote

Is there any named entity reconginition algorithm trained for the french language?

I'm also using Spacy models for french NER. You can re-train them to enhance the results. On the other hand, Google offers an api but it may get pretty expensive depending on the amount of text you ...
Abdaoui Amine's user avatar
1 vote

Entity Recognition in Stanford NLP using Python

From the following links, I understood that we can use a specific classifier by doing. Load the specific classifier: ...
nag's user avatar
  • 181
1 vote

How to extract entities from text using existing ontologies?

I think you are looking for Spacy, Polyglot and AllenNLP to find your NERs.
Syenix's user avatar
  • 359
1 vote

What is a lower bound on the vocabulary size for generating word/sentence embedding using word2vec or skip thought vectors?

It's not a straightforward question to answer as it is hard to compare the quality of two word2vec models with a meaningful metric. You could, of course, use the loss function, but that won't give ...
Escachator's user avatar
1 vote

Agglomerative Clustering without knowing number of clusters

If you don t know the number of clusters, i encourage you to look at those density based algorithm : Mean Shift, DBSCAN, OPTICS. They don t presume of the cluster number and are able to find random ...
KyBe's user avatar
  • 420
1 vote

Agglomerative Clustering without knowing number of clusters

A minimum cluster size will not generally be satisfiable in hierarchical clustering. Instead, you have to expect many clusters with just a single point. ELKI has some fairly interesting techniques to ...
Has QUIT--Anony-Mousse's user avatar
1 vote
Accepted

Tools to extract information

Open Calais is a free-to-use tool for entity recognition and relationship mapping. It's from Thompson-Reuters so may not be wholly suitable for technical language but worth a try. Has python bindings
julian's user avatar
  • 26
1 vote

Stanford NER is not properly extracting percentages

Looking at your code, I dont believe there is any mistake in it. The Stanford NER tagger is modelled by a Conditional Random Field Classifier and is hence bound to make mistakes. If you are able to ...
Himanshu Rai's user avatar
  • 1,848
1 vote
Accepted

which machine learning technique can be used?

I think these are the methods that you can try out (Please feel free to add more to this list): Highly precise with a little low recall is to use a dictionary with almost all possibilities (manual ...
Rahul Reddy Vemireddy's user avatar

Only top scored, non community-wiki answers of a minimum length are eligible