[Journey] 1 million UVs/month in 12 months using AI generated content. Let's do it!

Status
Not open for further replies.
how you can check this?
site:domain.com

On penalty it can give 1-10 results.
If you have much more pages / posts and your site is not fresh - you can suspect that.
 
https://towardsdatascience.com/keyword-extraction-python-tf-idf-textrank-topicrank-yake-bert-7405d51cd839
which method do you prefer to use sarte? and this is just for the main seo keyword right? where do you feed it after to generate the content?
I prefer Yake. i explained in detail before, you would have to go through the thread.
 
Thanks yeah I have several times you basically say it's proprietary
1. every keyword is stored in my db
2. every keyword has a snippet/article scraped and stored in the db
3. every keyword has a table "topical relevancy" where topics are scraped using yake
4. then i group the keyword who have the most similar topics in topical relevancy table
 
impressive, it suggests youtube videos via keyword or that was prexisting in scraped content/?
 
impressive, it suggests youtube videos via keyword or that was prexisting in scraped content/?
it suggests youtube vids based on the title of the article
 
How to check output text if it makes sense or not?

Libraries I tried like readability or language_tool_python either check for comprehension level or the grammatical errors, respectively.

I want to be able to check if text makes sense in the same way you read and know if a 2nd language/broken english person wrote something vs. a college/university graduate. How can python test for this?

In the specific case though, I want to be able to check both the quality of the content scraped as well as the paraphrasing output.
 
it suggests youtube vids based on the title of the article

hello ı want to ask you a question.
when i try to solve captcha its saying " Your computer is sending automated requests"
how can ı solve this problem ? can you give a hint how can i do it.
 
How to check output text if it makes sense or not?

Libraries I tried like readability or language_tool_python either check for comprehension level or the grammatical errors, respectively.

I want to be able to check if text makes sense in the same way you read and know if a 2nd language/broken english person wrote something vs. a college/university graduate. How can python test for this?

In the specific case though, I want to be able to check both the quality of the content scraped as well as the paraphrasing output.
That's not really possible easily. You need to have a model that outputs perfect English and only.then you can check for stuff like paraphrasing, readability.
hello ı want to ask you a question.
when i try to solve captcha its saying " Your computer is sending automated requests"
how can ı solve this problem ? can you give a hint how can i do it.
It's easier with Selenium.
 
yes i am using selenium. but its giving this error. " Your computer is sending automated requests"
are you doing several threads in parallel from 1 ip? How much sleep between clicks?
 
Just one thread. 10 secons.
https://prnt.sc/rSs-XCAhsHCs
spoof the user agent.
your IP might be poor. can you try some other IP like your mobile phone?

sleep between clicks for 0.5s
 
https://blog.google/products/search/how-were-improving-search-results-when-you-use-quotes/
Interesting. Any ideas how to incorporate this?
 
That's not really possible easily. You need to have a model that outputs perfect English and only.then you can check for stuff like paraphrasing, readability.

I explained it poorly. I didn't mean bad english but rather the paraphraser producing output that would be nonsense given the context of the sentence. It will make things that don't make sense when taken in context of a whole paragraph, whereas the original would have made sense.

Is it just a case of letting them ride (numbers game with the idea the rest of the article will make enough sense) or manual editing after - I am doing the latter right now as not going for huge output but rather just mimicking the type of article output a human whitehatter would do. I would prefer full automation though of course but I will not compromise on quality (below a certain level, I accept some concessions will be made for full automation).
 
Last edited:
Status
Not open for further replies.
Back
Top