General
-
Error 429
i am trying to give multiple requestes using for loop for extracting keywords but server is giving me error 429, how to resolve it?
2 votes -
Allow for URLs to be supplied to the /text/keywords endpoint
Currently it looks like the API does not allow for URLs to be supplied here (but please let me know if I'm wrong on this).
1 vote -
Unexpected low similarity between a text and its translation (to Spanish and back)
The original text, and its translation, only sharing 70%:
Original:
“Yes—for I shall be blamed, and rightly so!” cried the monk. “How can I best ensure that such a disaster will not occur? By extensive comments in the code? By extensive documentation outside the code? By providing a hundred unit tests that will fail if incorrect changes are made? By permeating the code with a hundred consistency checks—thus increasing its complexity even more? Or by remaining on this project until my brains turn to jelly?”Translated (through Spanish and back)
"Yes, because they will blame me, and with good reason!"…1 vote -
Is there any way to specify the number of keywords ???
i'm trying to extract keywords from text but i get 10 keywords all the time ! (i would like 20 keywords )
I m using the function getKeywordsForText(javaText); in java ...Thanks.
3 votesDear Sonja Ićeva,
Thank you for your suggestion regarding
providing a way to specify the number of keywords for our Retina API. At present our Retina API only supports 10 (ten) keywords, which we find is more than enough for the majority of use cases. One reason for limiting the number of keywords is to prevent the loss of context.We will keep your suggestion in mind.
Best regards from Vienna!
Mark
-
Actually, make Retina language addition automated!
Add ... Greek ?
1 vote -
Please develop some basic, written tutorials. Developers don't want to watch videos ... slow and i can't copy the code :)
For each major enviromnent/language please develop a written (printable / pdf) that describes how to get started and to write your first application using ypour API. Then supplement this with other examples illustrating simple but 'real' problems being solved through the software/API. - Thanks
1 vote -
error
Using a curl command I am able to detect Vietnamese, Korean, Thai & Japansese with http://api.cortical.io/rest/text/detect_language, but can only get a fingerprint for English & Vientnamese other languages tested returned an error.
"errordescription":"It is not possible to create a meaningful semantic fingerprint from the input text. The most likely causes are, that the input text was empty, too small, was not provided in UTF-8 format, or contained fragments of symbols or other non-alphabetic characters.","suggestedresolution":"Please try to provide either a larger text fragment and/or remove non-alphabetic characters from the input text (leaving punctuation characters in)."
Is Cortical able…
1 vote -
using language other than english
Except language detection, no api parameters indicates in which language we need to work.
i saw only "enassociative or en-synonymous" .And It seems we have to contact sales to access a retina french database.
So how can we try to validate interest ?
( i tried with default en.. but got only non significant words le,la, les, de, des ... which demonstrate no use of language )
is this product only trained for english ?1 vote -
keyword extraction from german text
i am using the fullclient rest-api within python. i tried to extract keywords from german text. unfortunately the results are quite poor. major problem is that stop words like "der", "die", "das" and so on are primarily returned. these are probably the most common terms in an usual text but also the least relevant. is it possible to optimize the language specific operations, e.g. train the algorithm myself?
1 vote -
Time series of texts (posts)
Tutorial on time series of texts
1 vote -
Scalable Comparison for Large Number of Documents
I have a large number of documents, of 500 to 2000 words, that I would like link to each other based on semantic similarity. The number of documents will increase rapidly so a process of comparing each document with all other documents will not scale.
As a solution I plan to retrieve a retina representation (Fingerprint) for each document from:
/text?retinaname=enassociativeThen find the 50 documents with the largest intersection of the values in the Fingerprint, then retrieve a representation of similarity with those 50 documents form:
compare?retinaname=enassociativeIs this the most efficient strategy ?
2 votes -
cURL support for compare?
I am trying to compare two texts using the curl API function, but I'm not certain how based on the example because it only says:
Request URL
http://api.cortical.io:80/rest/compare?retinaname=enassociativebut doesn't provide how we would input the two elements. I tried curl -X GET -H "api-key: xxxxx" "http://api.cortical.io:80/rest/compare?retinaname=enassociative&body=[{"term": "Pablo Picasso"},{"text": "Gustav Klimt was born in Baumgarten, near Vienna in Austria-Hungary, the second of seven children"}]"
but that did not work. Any help would be appreciated. Thanks.
3 votes -
Issues with Similarity Explorer
Human with Baby = 13% related but Shoes with Baby = 19% how exactly this is working i wanted to find out which is more related term to baby of course it's human not shoes.
1 vote -
Unexpected low similarity between a term and a text
I am using the compare api and ultimately I am trying to search for news articles that have to do with License agreements between companies. And in my first test I ran the following example but got a low 0.05 cosine similarity.
[
{
"term": "License Agreement"
},
{
"text": "Google and apple sign License Agreement to use a new search technology"
}]
1 vote -
Training Domain Knowledge on top of Retinas for detecting similar phrases
If we could use use topic modeler to define phrases in terms existing terms in Retina. Eg. Grease my hands = Corruption or Crime sub government. And after creating the signature train the retina's with that so next time if I have a sentence like "I will pass the deal just make sure Mr X hands are greased." Is it possible to do it in Amazon Image?
1 vote -
Unexpected response according to the context of a text!
With the interactive api I tried "/expressions/contexts" (http://api.cortical.io/Expression.htm#!/expressions/getContextsForExpressionpost1) with the following text input and I did not expect to get “cheese” as a result in the list.
{"text": "I was drinking beer with my friends at the Oktoberfest"
}
Do you have any suggestion to improve the result here? Maybe it would be awesome just to get words describing the context which are actually used in the text?2 votes -
Use of semantic fingerprints for a specific and isolated business domain?
Would it be possible to use the cortical.io Retina for a specific and isolated business domain that’s only accessible for one partner?
1 voteyes this would definitely be possible.
In the near future cortical.io’s technology will also be available as Amazon Machine Image via the Amazon Marketplace. Like this you would have your own cortical.io-Server instance.
Soon there will also be retinas in Spanish, French, German, Italian, Dutch, Swedish, Polish, Russian. You will be able to train specific domain language on top of the basic retinas.
- Don't see your idea?