We Keep Coding

How to detect features of a product in an english sentence - nlp

I am trying to detect features(eg.: screen, processing speed) of a product(eg.: mobile, respectively) in an english sentence. For this, my approach is that in a paragraph(that talks about the product) containing multiple sentences, the words( apart from words like pronouns or sentiment words like good, bad etc, which I store in a file) that appear most frequently are the features of that product and so I rank on the basis of their frequency and their distance with the sentiment words and take teh top n of them.
However, it is not very effective. Can anyone suggest some other and better approach for detecting the words which are features of a product?

There's been massive amount of research in this area. Start from reading Bing Liu's seminal work (Liu 2004, Liu 2005) in this area.
One popular technique is using Dependency Graph using Stanford CodeNLP.
You can make rules like a Noun (NN) connected to an Adjective (JJ) using nsubj dependency. 5-10 rules of this kind would be sufficient for a basic system.
State of the art in this area uses Sequence Tagging approach (CRF/HMM) for tagging each word whether it is an feature term or not. However you need good amount of labelled data for it. Check recent works in the area of Aspect Based Sentiment Analysis.
Resources for your help:
http://alt.qcri.org/semeval2015/task12/
http://www.aueb.gr/users/ion/docs/pavlopoulos_phd_thesis.pdf
http://www.aclweb.org/anthology/S14-2004

Identify an english word as a thing or product?

Write a program with the following objective -
be able to identify whether a word/phrase represents a thing/product. For example -
1) "A glove comprising at least an index finger receptacle, a middle finger receptacle.." <-Be able to identify glove as a thing/product.
2) "In a window regulator, especially for automobiles, in which the window is connected to a drive..." <- be able to identify regulator as a thing.
Doing this tells me that the text is talking about a thing/product. as a contrast, the following text talks about a process instead of a thing/product -> "An extrusion coating process for the production of flexible packaging films of nylon coated substrates consisting of the steps of..."
I have millions of such texts; hence, manually doing it is not feasible. So far, with the help of using NLTK + Python, I have been able to identify some specific cases which use very similar keywords. But I have not been able to do the same with the kinds mentioned in the examples above. Any help will be appreciated!

What you want to do is actually pretty difficult. It is a sort of (very specific) semantic labelling task. The possible solutions are:
create your own labelling algorithm, create training data, test, eval and finally tag your data
use an existing knowledge base (lexicon) to extract semantic labels for each target word
The first option is a complex research project in itself. Do it if you have the time and resources.
The second option will only give you the labels that are available in the knowledge base, and these might not match your wishes. I would give it a try with python, NLTK and Wordnet (interface already available), you might be able to use synset hypernyms for your problem.

This task is called named entity reconition problem.
EDIT: There is no clean definition of NER in NLP community, so one can say this is not NER task, but instance of more general sequence labeling problem. Anyway, there is still no tool that can do this out of the box.
Out of the box, Standford NLP can only recognize following types:
Recognizes named (PERSON, LOCATION, ORGANIZATION, MISC), numerical
(MONEY, NUMBER, ORDINAL, PERCENT), and temporal (DATE, TIME, DURATION,
SET) entities
so it is not suitable for solving this task. There are some commercial solutions that possible can do the job, they can be readily found by googling "product name named entity recognition", some of them offer free trial plans. I don't know any free ready to deploy solution.
Of course, you can create you own model by hand-annotating about 1000 or so product name containing sentences and training some classifier like Conditional Random Field classifier with some basic features (here is documentation page that explains how to that with stanford NLP). This solution should work reasonable well, while it won't be perfect of course (no system will be perfect but some solutions are better then others).
EDIT: This is complex task per se, but not that complex unless you want state-of-the art results. You can create reasonable good model in just 2-3 days. Here is (example) step-by-step instruction how to do this using open source tool:
Download CRF++ and look at provided examples, they are in a simple text format
Annotate you data in a similar manner
a OTHER
glove PRODUCT
comprising OTHER
...
and so on.
Spilt you annotated data into two files train (80%) and dev(20%)
use following baseline template features (paste in template file)
U02:%x[0,0]
U01:%x[-1,0]
U01:%x[-2,0]
U02:%x[0,0]
U03:%x[1,0]
U04:%x[2,0]
U05:%x[-1,0]/%x[0,0]
U06:%x[0,0]/%x[1,0]
4.Run
crf_learn template train.txt model
crf_test -m model dev.txt > result.txt
Look at result.txt. one column will contain your hand-labeled data and other - machine predicted labels. You can then compare these, compute accuracy etc. After that you can feed new unlabeled data into crf_test and get your labels.
As I said, this won't be perfect, but I will be very surprised if that won't be reasonable good (I actually solved very similar task not long ago) and certanly better just using few keywords/templates
ENDNOTE: this ignores many things and some best-practices in solving such tasks, won't be good for academic research, not 100% guaranteed to work, but still useful for this and many similar problems as relatively quick solution.

extracting the meaning of a sentence

Is there some java library that helps extract the content of a sentence/paragraph?
Essentially what I need to do is get a context of what is being said (such as whether the sentence is providing a positive or negative point and that sort of thing).
I don't know of such a system and have been looking around and have not been able to find anything useful. does anyone know of something that might help with this?
thanks

Use GATE (https://gate.ac.uk/), a NLP & Machine Learning tool.
You can use ANNIE for splitting sentences and POS tagging.
You have to prepare training dataset, with sentiments already annotated manually and then use Batch Learning plugin to predict sentiment for training documents.
Step-by-step tutorial for this: https://gate.ac.uk/sale/talks/gate-course-may10/track-3/module-11-ml-adv/module-11-sentiment.pdf
And the example talked about in the pdf: https://gate.ac.uk/sale/talks/gate-course-may10/track-3/module-11-ml-adv/module-11-sentiment.zip

simple sentiment analysis with java

I am very new to Sentiment analysis. How can I judge if a given word or sentence is positive or negative. I have to implement it with java. I tried to read something like lingpipe, rapidminer tutorial, but I do not understand. In their examples they use a lot of data. In my case I do not have much data. All I have is a word or a sentence, lets say. I tried to read the questions from stackoverflow too. But they do not help me much.
Thanks in advance.

Computers don't know about a human thing like sentiment unless they learn it from examples that a human has labeled as positive or negative.
The goal of Machine Learning is in fact to make the most informed decision about a new example based on the empirical data of previous examples. Statistically, the more data, the better.
To "judge" the sentiment of a sentence, you'll need to have trained a model or classifier on some sentences labeled for sentiment. The classifier takes an unlabeled sentence as input and outputs a label: positive or negative.
First get training examples. I'm sure you can find some labeled sentiment data in the public domain. One of the best data set repositories is the UCI KDD Archive. You may then train a classifier on the data to judge new examples. There are a host of learning algorithm resources available. My favorites are jBoost, which can output a classifier as Java code, and Rapidminer, which is better for visual analysis.

You could use an existing web-service which is trained from prior data. For example:
Chatterbox Sentiment Detection API
Which has libraries for Java & Android.
(Disclosure: I work for the company that builds this API)

This is not really programming related (neuro-linguistic programming is not programming), and in general there is no reliable solution.
My best idea is to make it work like Google "Pigeon"Rank, i.e. collect words and sentences, and then collect human feedback whether they are positive or negative, and then use Bayesian matching with this data.

Your can try to use Wordnet for searching word's Semantic Orientation based on "distance" calculation between your word and "good" or "bad" words.Shorter distance will give you word's SO. Results seems will be a bit weak but not a lot of data(or time) is necessary for this approach.

Build a natural language model that fixes misspellings

What are books about how to build a natural language parsing program like this:
input: I got to TALL you
output: I got to TELL you
input: Big RAT box
output: Big RED box
in: hoo un thum zend three
out: one thousand three
It must have the language model that allows to predict what words are misspelled !
What are the best books on how to build such a tool??
p.s. Are there free webservices to spell-check? From Google maybe?..

Peter Norvig has written a terrific spell checker. Maybe that can help you.

You have at least three options
You can write a program which understands the language (i.e. what a word means). This is a topic for research today. Expect the first results when you can buy a computer which is fast enough to run such a program (which is probably in 10 years when computers have become 1000 times faster than today).
Use a huge corpus (text documents) to train a Hidden Marcov Model.
Use a huge corpus and generate statistics about quadruplets n-grams, i.e. how often a tuple of N words appears. I don't have a link handy for this but the idea is that some words always appear in the context of other words. So when you parse your text into 4-grams and look them up in your database and you can't find one, chances are that there is something wrong with the current tuple. The next step is to find all possible matches (other 4-grams which have a small soundex or similar distance to the current one) and try the one with the highest frequency.
Google has this data for quite a few languages and you might find more in Google labs about this.
[EDIT] After some googling, I finally found the link: On this page, you can buy English 1- to 5-grams which Google collected over the whole Internet on 6 DVDs.
Googling for "google spelling statistics n-grams" will also turn up some interesting links.

soundex (wiki) is one option

There are quite a few Java libraries for natural language processing that would help you implement a spelling corrector. But you asked about a book. Foundations of Statistical Natural Language Processing by Christopher D. Manning and Hinrich Schütze looks like a good option. The first author is a Stanford Professor leading a group that does natural language processing and developing Java libraries and NLP resources that many people use.

In Dev Days London, Michael Sparks presented a Python script coded exactly for that. It was surprisingly very simple! See if you can find in Google. Maybe somebody here will have the link.