I would like to know whether there is a proper term to describe "diffing" of / obtaining the delta between multiple files or data structures, such that the resulting "diff" contains first a description of the parts common to all files/structures, then descriptions of how this "base" file/structure must be modified to obtain the individual ones, ideally in a hierarchical fashion if some files/structures are more similar to each other than others.
There are some questions and answers about how to do this with certain tools (e.g. DIFF utility works for 2 files. How to compare more than 2 files at a time?), but as I want to do this for a specific type of data structure (namely JSON), I'm at a loss as to what I should even search for.
This type of problem seems to me like it should be common enough to have a name such as "hierarchical diff" (which however seems to be reserved for 2-way diffs on hierarchical data structures), "commonality finding", or something like that.
I guess a related concept about hierarchical ordering of commonalities and differences is formal concept analysis, but this operates on sets of properties rather than hierarchical data structures and won't help me much.
There are multiple valid denominations :
Data comparison (or Sequence comparison)
Delta encoding
Delta compression (or Differential compression)
Algorithms:
An O(ND) Difference Algorithm and Its Variations (Eugene Myer)
A technique for isolating differences between files (Paul Heckel)
The String-to-String Correction Problem with Block Moves (Walter Tichy)
Good Wikipedia links
Longest common subsequence problem
Comparison of file comparison tools
Diff Unix Utility
Some implementations
diff-match-patch (Neil Fraser - Google)
jsdifflib
jsondiffpatch
Related
I'm trying to learn about scanning and reading encoded data in barcodes both 1D and 2D. ISO/IEC 15418 seems to detail very closely the data I am interested in reading. Unfortunately, the specification is not good at giving full examples of what the specification looks like in practice.
Things like Record Separators and Group Separators (see ASCII characters 29 and 30) appear out of nowhere without any definition in the ANSI specification.
ANS MH10.8.2-2016 (ISO/IEC 15418) PDF
Also relevant: ANS MH10.8.17-2017 (ISO/IEC 15434) PDF
So far, our way of explaining this is by scanning barcodes (data matrixes seem to provide data like the specification shows most commonly), reading the data and checking the document, slowly identifying patterns around how the data is structured. This specification seems to only lightly touch on the structure of data, but goes into exact detail about constituent parts.
I understand that my questions are somewhat broad and unspecific, but I can barely find anything about this specification to begin with. There's barely anything on the entirety of StackOverflow.
General Questions
Where can I find full examples and explanations of what this specification looks like in practice?
Are there any publicly available parsers or APIs surrounding this?
Where should I look for more information, ask questions etc. about this specification and ones like it?
Specific Questions
When scanning barcodes, data matrixes and QR codes, how is one supposed to easily differentiate data stored in this standard versus raw text with no particular standard applied?
a. We assume that there must be a industry standard for doing this besides "check if it kinda looks like one based on what we know".
Currently, the barcodes I am scanning seem to primarily use Data Identifiers, not Application Identifiers. However, online, I did manage to find an example of someone using an Application Identifier, and it very closely resembled a Data Identifier structure like I scanned. Assuming there will be ambiguity, what is the difference between the two?
Do 1D barcodes actually even store this kind of data? So far, I cannot remember if I've scanned a 1D barcode storing more than simple 'text'. Only Data Matrixes provide the raw encoded data from the specifiation.
I have this classification scenario below in which Im getting a very low F1, precision, recall and other metrics.
Target is multiclass (about ~200 classes) which is highly imbalanced
I only use company names as classifier (mostly 1-2 words which have max of 8 words), no other fields (like description, etc.)
Training data ~ 100k+ records
Preprocessing: numeric and special characters and stopwords removal
I have very low resources for processing (thats why when I try to use oversampling techniques like smote, distance_smote for multiclass, etc., I always get memory error)
Tried using different vectorization/embedding/tokenizer like word2vec, tfidf, fasttext, bert, roberta, etc. but to no avail
Tried using (and fine-tuning) different algorithms (networks, svm, trees, boosting, etc.) but also getting low scores.
I also did cost-sensitive learning (using class weights) but it only decreased my scores.
Tried all options that I know but scores are not increasing. Can you recommend other options here or do you think any part of the process that may be wrong/discarded? Thank you!
Distribution of target labels:
Sample observations
There is essentially no way to know that 'Exxon' is an oil company, and 'Apple' a computer company, and 'McDonalds' a fast-food chain, just from their company names.
Even if you have a list of every other company in the world, by name and type, that's not enough to make the deduction for these last 3. Only other outside info – like a few sentences about them, or other data – could classify them.
In fact, while company names sometimes describe their exact field-of-commerce, often they're totally arbitrary, as that gives them more freedom to range over many products/services, or create their own unique associations with the name (aka branding).
So I strongly suspect your (unshown) names & (unshown) labels are just too arbitrary for the data you're using to get very good at the task you're attempting.
Is there a real-world situation where someone will only have a company name – no other info, or research options – and benefit from correctly guessing the class? If so, more specifics about the situation might help generate more specific tactical recommendations. But mainly such recommendations will be: get richer data about the targets of the classification.
You might squeeze a little more out of vague trends in corporate naming via better preprocessing/feature-extraction. You may want to keep numbers, special-characters, & punctuation in some form, as they might include extra slight hints. Using subwords (character n-grams) might also reveal some shared word-roots used even in made-up names.
I have a lot of texts (millions), ranging from 100 to 4000 words. The texts are formatted as written work, with punctuation and grammar. Everything is in English.
The problem is simple: How to extract every WikiData entity from a given text?
An entity is defined as every noun, proper or regular. I.e., names of people, organizations, locations and things like chair, potatoes etc.
So far I've tried the following:
Tokenize the text with OpenNLP, and use the pre-trained models to extract people, location, organization and regular nouns.
Apply Porter Stemming where applicable.
Match all extracted nouns with the wmflabs-API to retrieve a potential WikiData ID.
This works, but I feel like I can do better. One obvious improvement would be to cache the relevant pieces of WikiData locally, which I plan on doing. However, before I do that, I want to check if there are other solutions.
Suggestions?
I tagged the question Scala because I'm using Spark for the task.
Some suggestions:
consider Stanford NER in comparison to OpenNLP to see how it compares on your corpus
I wonder at the value of stemming for most entity names
I suspect you might be losing information by dividing the task into discrete stages
although Wikidata is new, the task isn't, so you might look at papers for Freebase|DBpedia|Wikipedia entity recognition|disambiguation
In particular, DBpedia Spotlight is one system designed for exactly this task.
http://static.googleusercontent.com/media/research.google.com/en//pubs/archive/38389.pdf
http://ceur-ws.org/Vol-1057/Nebhi_LD4IE2013.pdf
I am working with some really large databases of newspaper articles, I have them in a MySQL database, and I can query them all.
I am now searching for ways to help me tag these articles with somewhat descriptive tags.
All these articles is accessible from a URL that looks like this:
http://web.site/CATEGORY/this-is-the-title-slug
So at least I can use the category to figure what type of content that we are working with. However, I also want to tag based on the article-text.
My initial approach was doing this:
Get all articles
Get all words, remove all punctuation, split by space, and count them by occurrence
Analyze them, and filter common non-descriptive words out like "them", "I", "this", "these", "their" etc.
When all the common words was filtered out, the only thing left is words that is tag-worthy.
But this turned out to be a rather manual task, and not a very pretty or helpful approach.
This also suffered from the problem of words or names that are split by space, for example if 1.000 articles contains the name "John Doe", and 1.000 articles contains the name of "John Hanson", I would only get the word "John" out of it, not his first name, and last name.
Automatically tagging articles is really a research problem and you can spend a lot of time re-inventing the wheel when others have already done much of the work. I'd advise using one of the existing natural language processing toolkits like NLTK.
To get started, I would suggest looking at implementing a proper Tokeniser (much better than splitting by whitespace), and then take a look at Chunking and Stemming algorithms.
You might also want to count frequencies for n-grams, i.e. a sequences of words, instead of individual words. This would take care of "words split by a space". Toolkits like NLTK have functions in-built for this.
Finally, as you iteratively improve your algorithm, you might want to train on a random subset of the database and then try how the algorithm tags the remaining set of articles to see how well it works.
You should use a metric such as tf-idf to get the tags out:
Count the frequency of each term per document. This is the term frequency, tf(t, D). The more often a term occurs in the document D, the more important it is for D.
Count, per term, the number of documents the term appears in. This is the document frequency, df(t). The higher df, the less the term discriminates among your documents and the less interesting it is.
Divide tf by the log of df: tfidf(t, D) = tf(t, D) / log(df(D) + 1).
For each document, declare the top k terms by their tf-idf score to be the tags for that document.
Various implementations of tf-idf are available; for Java and .NET, there's Lucene, for Python there's scikits.learn.
If you want to do better than this, use language models. That requires some knowledge of probability theory.
Take a look at Kea. It's an open source tool for extracting keyphrases from text documents.
Your problem has also been discussed many times at http://metaoptimize.com/qa:
http://metaoptimize.com/qa/questions/1527/what-are-some-good-toolkits-to-get-lda-like-tagging-of-my-documents
http://metaoptimize.com/qa/questions/1060/tag-analysis-for-document-recommendation
If I understand your question correctly, you'd like to group the articles into similarity classes. For example, you might assign article 1 to 'Sports', article 2 to 'Politics', and so on. Or if your classes are much finer-grained, the same articles might be assigned to 'Dallas Mavericks' and 'GOP Presidential Race'.
This falls under the general category of 'clustering' algorithms. There are many possible choices of such algorithms, but this is an active area of research (meaning it is not a solved problem, and thus none of the algorithms are likely to perform quite as well as you'd like).
I'd recommend you look at Latent Direchlet Allocation (http://en.wikipedia.org/wiki/Latent_Dirichlet_allocation) or 'LDA'. I don't have personal experience with any of the LDA implementations available, so I can't recommend a specific system (perhaps others more knowledgeable than I might be able to recommend a user-friendly implementation).
You might also consider the agglomerative clustering implementations available in LingPipe (see http://alias-i.com/lingpipe/demos/tutorial/cluster/read-me.html), although I suspect an LDA implementation might prove somewhat more reliable.
A couple questions to consider while you're looking at clustering systems:
Do you want to allow fractional class membership - e.g. consider an article discussing the economic outlook and its potential effect on the presidential race; can that document belong partly to the 'economy' cluster and partly to the 'election' cluster? Some clustering algorithms allow partial class assignment and some do not
Do you want to create a set of classes manually (i.e., list out 'economy', 'sports', ...), or do you prefer to learn the set of classes from the data? Manual class labels may require more supervision (manual intervention), but if you choose to learn from the data, the 'labels' will likely not be meaningful to a human (e.g., class 1, class 2, etc.), and even the contents of the classes may not be terribly informative. That is, the learning algorithm will find similarities and cluster documents it considers similar, but the resulting clusters may not match your idea of what a 'good' class should contain.
Your approach seems sensible and there are two ways you can improve the tagging.
Use a known list of keywords/phrases for your tagging and if the count of the instances of this word/phrase is greater than a threshold (probably based on the length of the article) then include the tag.
Use a part of speech tagging algorithm to help reduce the article into a sensible set of phrases and use a sensible method to extract tags out of this. Once you have the articles reduced using such an algorithm, you would be able to identify some good candidate words/phrases to use in your keyword/phrase list for method 1.
If the content is an image or video, please check out the following blog article:
http://scottge.net/2015/06/30/automatic-image-and-video-tagging/
There are basically two approaches to automatically extract keywords from images and videos.
Multiple Instance Learning (MIL)
Deep Neural Networks (DNN), Recurrent Neural Networks (RNN), and the variants
In the above blog article, I list the latest research papers to illustrate the solutions. Some of them even include demo site and source code.
If the content is a large text document, please check out this blog article:
Best Key Phrase Extraction APIs in the Market
http://scottge.net/2015/06/13/best-key-phrase-extraction-apis-in-the-market/
Thanks, Scott
Assuming you have pre-defined set of tags, you can use the Elasticsearch Percolator API like this answer suggests:
Elasticsearch - use a "tags" index to discover all tags in a given string
Are you talking about the name-entity recognition ? if so, Anupam Jain is right. it;s research problem with using deep learning & CRF. In 2017, the name-entity recognition problem is force on semi-surprise learning technology.
The below link is related ner of paper:
http://ai2-website.s3.amazonaws.com/publications/semi-supervised-sequence.pdf
Also, The below link is key-phase extraction on twitter:
http://jkx.fudan.edu.cn/~qzhang/paper/keyphrase.emnlp2016.pdf
I want to build a generic Perl module for handling and analysing biomedical character separated datasets and which can, most certain, be used on any kind of datasets that contain a mixture of categorical (A,B,C,..) and continuous (1.2,3,881..) and identifier (XXX1,XXX2...). The plan is to have people initialize the module and then use some arguments to point to the data file(s), the place were the analysis reports should be placed and the structure of the data.
By structure of data I mean which variable is in which place and its name/type. And this is where I need some enlightenment. I am baffled how to do this in a clean way. Obviously, having people create a simple schema file, be it XML or some other format would be the cleanest but maybe not all people enjoy doing something like this.
The solutions I can think of are:
Create a configuration file in XML or similar and with a prespecified format.
Pass the information during initialization of the module.
Use the first row of the data as headers and try to guess types (ouch)
Surely there must be a "canonical" way of doing this that is also usable and efficient.
This doesn't answer your question directly, but have you checked CPAN? It might have the module you need already. If not, it might have similar modules -- related either to biomedical data or simply to delimited data handling -- that you can mine for good ideas, both concerning formats for metadata and your module's API.
Any of the approaches you've listed could make sense. It all depends on how complex the data structures and their definitions are. What will make something like this useful to people is whether it saves them time and effort. So, your decision will have to be answered based on what approach will best satisfy the need to make:
use of the module easy
reuse of data definitions easy
the data definition language sufficiently expressive to describe all known use cases
the data definition language sufficiently simple that an infrequent user can spend minimal time with the docs before getting real work done.
For example, if I just need to enter the names of the columns and their types (and there are only 4 well defined types), doing this each time in a script isn't too bad. Unless I have 350 columns to deal with in every file.
However, if large, complicated structure definitions are common, then a more modular reuse oriented approach is better.
If your data description language is difficult to work with, you can mitigate the issue a bit by providing a configuration tool that allows one to create and edit data schemes.
rx might be worth looking at, as well as the Data::Rx module on the CPAN. It provides schema checking for JSON, but there is nothing inherent in the model that makes it JSON-only.