I want to know the effective algorithms/data structures to identify the below information in streaming data.
Consider a real-time streaming data like twitter. I am mainly interested in the below queries rather than storing the actual data.
I need my queries to run on actual data but not any of the duplicates.
As I am not interested in storing the complete data, it will be difficult for me to identify the duplicate posts. However, I can hash all the posts and check against them. But I would like to identify near duplicate posts also. How can I achieve this.
Identify the top k topics being discussed by the users.
I want to identify the top topics being discussed by users. I don't want the top frequency words as shown by twitter. Instead I want to give some high level topic name of the most frequent words.
I would like my system to be real-time. I mean, my system should be able to handle any amount of traffic.
I can think of map reduce approach but I am not sure how to handle synchronization issues. For example, duplicate posts can reach different nodes and both of them could store them in the index.
In a typical news source, one will be removing any stop words in the data. In my system I would like to update my stop words list by identifying top frequent words across a wide range of topics.
What will be effective algorithm/data structure to achieve this.
I would like to store the topics over a period of time to retrieve interesting patterns in the data. Say, friday evening everyone wants to go to a movie. what will be the efficient way to store this data.
I am thinking of storing it in hadoop distributed file system, but over a period of time, these indexes become so large that I/O will be my major bottleneck.
Consider multi-lingual data from tweets around the world. How can I identify similar topics being discussed across a geographical area?
There are 2 problems here. One is identifying the language being used. It can be identified based on the person tweeting. But this information might affect the privacy of the users. Other idea, could be running it through a training algorithm. What is the best method currently followed for this. Other problem is actually looking up the word in a dictionary and associating it to common intermediate language like say english. How to take care of word sense disambiguation like a same word being used in different contests.
Identify the word boundaries
One possibility is to use some kind of training algorithm. But what is the best approach followed. This is some way similar to word sense disambiguation, because you will be able to identify word boundaries based on the actual sentence.
I am thinking of developing a prototype and evaluating the system rather than the concrete implementation. I think its not possible to scrap the real-time twitter data. I am thinking this approach can be tested on some data freely available online. Any ideas, where I can get this data.
Your feedback is appreciated.
Thanks for your time.
-- Bala
There are a couple different questions buried in here. I can't understand all that you're asking, but here's a the big one as I understand it: You want to categorize messages by topic. You also want to remove duplicates.
Removing duplicates is (relatively) easy. To remove "near" duplicates, you could first remove uninteresting parts from your data. You could start by removing capitalization and punctuation. You could also remove the most common words. Then you could add the resulting message to a Bloom filter. Hashing isn't good enough for Twitter, as the hashed messages wouldn't be much smaller than the full messages. You'd end up with a hash that doesn't fit in memory. That's why you'd use a Bloom filter instead. It might have to be a very large Bloom filter, but it will still be smaller than the hash table.
The other part is a difficult categorization problem. You probably do not want to write this part yourself. There are a number of libraries and programs available for categorization, but it might be hard to find one that fits your needs. An example is the Vowpal Wabbit project, which is a fast online algorithm for categorization. However, it only works on one category at a time. For multiple categories, you would have to run multiple copies and train them separately.
Identifying the language sounds less difficult. Don't try to do something smart like "training", instead put the most common words from each language in a dictionary. For each message, use the language whose words appeared most frequently.
If you want the algorithm to come up with categories on its own, good luck.
I'm not really sure if I'm answering your main question, but you could determine the similarity of two messages by calculating the Levenshtein distance between them. You can think of this as the "edit difference" between two strings (I.E., how many edits would need to be made to one, to convert it to the other).
Hello we have created a very similar demo using api.cortical.io functionality.
There you can create semantic fingerprints of each tweet. (you could also extract the top most keywords or some similar terms, that don't need to actually be part of the tweet).
We have used the fingerprints to filter the twitter stream based on content.
On twistiller.com you can see the result. The public 1% twitter stream is monitored for four different topic areas.
Related
I'm currently having a lot of data (subtitles for films) which contains "compound words" in english.
There is no certain pattern, but in some places words are merged. E.g.:
You got to keep going. They're always looking **forsomething** under the seats.
I don't have specific constraints on language/platform, however there is really lots and lots of data and I won't be able to check results properly. So I need kind of highly-precise solution here.
So what are basically the ways to attack this kind of data?
I have to manually go through a long list of terms (~3500) which have been entered by users through the years. Beside other things, I want to reduce the list by looking for synonyms, typos and alternate spellings.
My work will be much easier if I can group the list into clusters of possible typos before starting. I was imagining to use some metric which can calculate the similarity to a term, e.g. in percent, and then cluster everything which has a similarity higher than some threshold. As I am going through it manually anyway, I don't mind a high failure rate, if it can keep the whole thing simple.
Ideally, there exists some easily available library to do this for me, implemented by people who know what they are doing. If there is no such, then at least one calculating a similarity metric for a pair of strings would be great, I can manage the clustering myself.
If this is not available either, do you know of a good algorithm which is simple to implement? I was first thinking a Hamming distance divided by word length will be a good metric, but noticed that while it will catch swapped letters, it won't handle deletions and insertions well (ptgs-1 will be caught as very similar to ptgs/1, but hematopoiesis won't be caught as very similar to haematopoiesis).
As for the requirements on the library/algorithm: it has to rely completely on spelling. I know that the usual NLP libraries don't work this way, but
there is no full text available for it to consider context.
it can't use a dictionary corpus of words, because the terms are far outside of any everyday language, frequently abbreviations of highly specialized terms.
Finally, I am most familiar with C# as a programming language, and I already have a C# pseudoscript which does some preliminary cleanup. If there is no one-step solution (feed list in, get grouped list out), I will prefer a library I can call from within a .NET program.
The whole thing should be relatively quick to learn for somebody with almost no previous knowledge in information retrieval. This will save me maybe 5-6 hours of manual work, and I don't want to spend more time than that in setting up an automated solution. OK, maybe up to 50% longer if I get the chance to learn something awesome :)
The question: What should I use, a library, or an algorithm? Which ones should I consider? If what I need is a library, how do I recognize one which is capable of delivering results based on spelling alone, as opposed to relying on context or dictionary use?
edit To clarify, I am not looking for actual semantic relatedness the way search or recommendation engines need it. I need to catch typos. So, I am looking for a metric by which mouse and rodent have zero similarity, but mouse and house have a very high similarity. And I am afraid that tools like Lucene use a metric which gets these two examples wrong (for my purposes).
Basically you are looking to cluster terms according to Semantic Relatedness.
One (hard) way to do it is following Markovitch and Gabrilovitch approach.
A quicker way will be consisting of the following steps:
download wikipedia dump and an open source Information Retrieval library such as Lucene (or Lucene.NET).
Index the files.
Search each term in the index - and get a vector - denoting how relevant the term (the query) is for each document. Note that this will be a vector of size |D|, where |D| is the total number of documents in the collection.
Cluster your vectors in any clustering algorithm. Each vector represents one term from your initial list.
If you are interested only in "visual" similarity (words are written similar to each other) then you can settle for levenshtein distance, but it won't be able to give you semantic relatedness of terms.For example, you won't be able to relate between "fall" and "autumn".
I have a large set of data (several hundred thousand records) that are unique entries in a CSV. These entries are essentially products that are being listed in a store from a vendor that offers these products. The problem is that while they offer us rights to copy these verbatim or to change wording, I don't want to list them verbatim obviously since Google will slap the ranking for having "duplicate" content. And then, also obviously, manually editing 500,000 items would take a ridiculous amount of time.
The solution, it would seem, would be to leverage fuzzy logic that would take certain phraseology and transform it to something different that would not then be penalized by Google. I have hitherto been unable to find any real library to address this or a solid solution that addresses such a situation.
I am thinking through my own algorithms to perhaps accomplish this, but I hate to reinvent the wheel or, worse, be beaten down by the big G after a failed attempt.
My idea is to simply search for various phrases and words (sans stop words) and then essentially map those to phrases and words that can be randomly inserted that still have equivalent meaning, but enough substance to hopefully not cause a deranking situation.
A solution for Ruby would be optimal, but absolutely not necessary as any language can be used.
Are there any existing algorithms, theories or implementations of a similar scenario that could be used to model or solve such a scenario?
Given a few words of input, I want to have a utility that will return a diverse set of relevant terms, phrases, or concepts. A caveat is that it would need to have a large graph of terms to begin with, or else the feature would not be very useful.
For example, submitting "baseball" would return
["shortstop", "Babe Ruth", "foul ball", "steroids", ... ]
Google Sets is the best example I can find of this kind of feature, but I can't use it since they have no public API (and I wont go against their TOS). Also, single-word input doesn't garner a very diverse set of results. I'm looking for a solution that goes off on tangents.
The closest I've experimented with is using WikiPedia's API to search Categories and Backlinks, but there's no way to directly sort those results by "relevance" or "popularity". Without that, the suggestion list is massive and all over the place, which is not immediately useful and very hard to whittle down.
Using A Thesaurus could also work minimally, but that would leave out any proper nouns or tangentially relevant terms (like any of the results listed above).
I would happily reuse an open service, if one exists, but I haven't found anything sufficient.
I'm looking for either a way to implement this either in-house with a decently-populated starting set, or reuse a free service that offers this.
Have a solution? Thanks ahead of time!
UPDATE: Thank you for the incredibly dense & informative answers. I'll choose a winning answer in 6 to 12 months, when I'll hopefully understand what you've all suggested =)
You might be interested in WordNet. It takes a bit of linguistic knowledge to understand the API, but basically the system is a database of meaning-based links between English words, which is more or less what you're searching for. I'm sure I can dig up more information if you want it.
Peter Norvig (director of research at Google) spoke about how they do this at Google (specifically mentioning Google Sets) in a Facebook Tech Talk. The idea is that a relatively simple algorithm on a huge dataset (e.g. the entire web) is much better than a complicated algorithm on a small data set.
You could look at Google's n-gram collection as a starting point. You'd start to see what concepts are grouped together. Norvig hinted that internally Google has up to 7-grams for use in things like Google Translate.
If you're more ambitious, you could download all of Wikipedia's articles in the language you desire and create your own n-gram database.
The problem is even more complicated if you just have a single word; check out this recent thesis for more details on word sense disambiguation.
It's not an easy problem, but it is useful as you mentioned. In the end, I think you'll find that a really successful implementation will have a relatively simple algorithm and a whole lot of data.
Take a look at the following two papers:
Clustering User Queries of a Search Engine [pdf]
Topic Detection by Clustering Keywords [pdf]
Here is my attempt at a very simplified explanation:
If we have a database of past user queries, we can define a similarity function between two queries. For example: number of words in common. Now for each query in our database, we compute its similarity with each other query, and remember the k most similar queries. The non-overlapping words from these can be returned as "related terms".
We can also take this approach with a database of documents containing information users might be searching for. We can define the similarity between two search terms as the number of documents containing both divided by the number of documents containing either. To decide which terms to test, we can scan the documents and throw out words that are either too common ('and', 'the', etc.) or that are too obscure.
If our data permits, then we could see which queries led users to choosing which results, instead of comparing documents by content. For example if we had data that showed us that users searching for "Celtics" and "Lakers" both ended up clicking on espn.com, then we could call these related terms.
If you're starting from scratch with no data about past user queries, then you can try Wikipedia, or the Bag of Words dataset as a database of documents. If you are looking for a database of user search terms and results, and if you are feeling adventurous, then you can take a look at the AOL Search Data.
I'm using Gensim's excellent library to compute similarity queries on a corpus using LSI. However, I have a distinct feeling that the results could be better, and I'm trying to figure out whether I can adjust the corpus itself in order to improve the results.
I have a certain amount of control over how to split the documents. My original data has a lot of very short documents (mean length is 12 words in a document, but there exist documents that are 1-2 words long...), and there are a few logical ways to concatenate several documents into one. The problem is that I don't know whether it's worth doing this or not (and if so, to what extent). I can't find any material addressing this question, but only regarding the size of the corpus, and the size of the vocabulary. I assume this is because, at the end of the day, the size of a document is bounded by the size of the vocabulary. But I'm sure there are still some general guidelines that could help with this decision.
What is considered a document that is too short? What is too long? (I assume the latter is a function of |V|, but the former could easily be a constant value.)
Does anyone have experience with this? Can anyone point me in the direction of any papers/blog posts/research that address this question? Much appreciated!
Edited to add:
Regarding the strategy for grouping documents - each document is a text message sent between two parties. The potential grouping is based on this, where I can also take into consideration the time at which the messages were sent. Meaning, I could group all the messages sent between A and B within a certain hour, or on a certain day, or simply group all the messages between the two. I can also decide on a minimum or maximum number of messages grouped together, but that is exactly what my question is about - how do I know what the ideal length is?
Looking at number of words per document does not seem to me to be the correct approach. LSI/LSA is all about capturing the underlying semantics of the documents by detecting common co-occurrences.
You may want to read:
LSI: Probabilistic Analysis
Latent Semantic Analysis (particularly section 3.2)
A valid excerpt from 2:
An important feature of LSI is that it makes no assumptions
about a particular generative model behind the data. Whether
the distribution of terms in the corpus is “Gaussian”, Poisson, or
some other has no bearing on the effectiveness of this technique, at
least with respect to its mathematical underpinnings. Thus, it is
incorrect to say that use of LSI requires assuming that the attribute
values are normally distributed.
The thing I would be more concerned is if the short documents share similar co-occurring terms that will allow LSI to form an appropriate topic grouping all of those documents that for a human share the same subject. This can be hardly done automatically (maybe with a WordNet / ontology) by substituting rare terms with more frequent and general ones. But this is a very long shot requiring further research.
More specific answer on heuristic:
My best bet would be to treat conversations as your documents. So the grouping would be on the time proximity of the exchanged messages. Anything up to a few minutes (a quarter?) I would group together. There may be false positives though (strongly depending on the actual contents of your dataset). As with any hyper-parameter in NLP - your mileage will vary... so it is worth doing a few experiments.
Short documents are indeed a challenge when it comes to applying LDA, since the estimates for the word co-occurrence statistics are significantly worse for short documents (sparse data). One way to alleviate this issue is, as you mentioned, to somehow aggregate multiple short texts into one longer document by some heuristic measure.
One particularity nice test-case for this situation is topic modeling Twitter data, since it's limited by definition to 140 characters. In Empirical Study of Topic Modeling in Twitter (Hong et al, 2010), the authors argue that
Training a standard topic model on aggregated user messages leads to a
faster training process and better quality.
However, they also mention that different aggregation methods lead to different results:
Topics learned by using different aggregation strategies of
the data are substantially different from each other.
My recommendations:
If you are using your own heuristic for aggregating short messages into longer documents, make sure to experiment with different aggregation techniques (potentially all the "sensical" ones)
Consider using a "heuristic-free" LDA variant that is better tailored for short messages, e.g, Unsupervised Topic Modeling for Short Texts Using Distributed
Representations of Words