Help with my Stop words script, Please

[john278]

Hi,

I have some spam submissions happening on my site and I want them to stop. To do this I have decided to check for words in a string and if they match then it will not submit that data. The thing is I am not quite sure how to put this together.

I think I have the structure write or nearly write, but then using the functions to make it all work is a bit dazling to me.

$stopwords = "games,gambling,about";

$text = "nothing about it most that most don't know which ones don and just discovered normally thinks they can make quick buck two well's not";

$wordex = preg_match();

if($wordex)

    {

echo "Spam";

}else{

echo "No Spam";

    } 

What I want in the above script is for it to loop through the $stopwords list while trying to match the words with the $text string. If there is a match with any words in the $stopwords list then it will echo "Spam" and if there is no match it will echo "No Spam".

How can I modify my script to do that.

Thanks!

[matsko]

so let me get this straight, you want to create a script that involves a list of keywords (in a string and separated by commas) and then you want them to be compared to an inputted text to see if the text contains any of them...

Well you could do this...

$words = "one,two,three,four,five,six";

//Convert Keywords to array
$words = explode(',',$words);

//Spam Bool
$spam = false;

//Check Words
foreach($words as $w)
{

  if(stripos(&$text,$w)!=-1)
  { 
    $spam = true;
    break;
  }

}

//Check Bool
if($spam==true)
echo 'spam'; 

Adding URL websites for keywords would be more effective then just keywords...

When it comes down to preventing spam on a website, a simple keyword system wont do the job. You see having text like G.A.M.E.S or game(s) wont be considered spam (since its not within the keyword list). Also adding more keywords will only slow down the website. So to prevent spam you need to take a look at what kind of messages your spam users are sending.

Most of the time when the average spam user sends a message he or she will send the same message to multiple users, or maybe just alter a few things in the message (name, user, etc...). By using strcmp() you can compare the templated spam message (the old message) to the message that was inputted. By setting a ballpark percentage (say 80&#37 then if the two messages are greater than 80% the same then they will be considered spam. So in order to do this, make a database system that flags the id of the last message that was sent by the user. Then once he/she sends another message it will compare the old message (fetched from the database via the flagged id) with the new inputted message. If both messages are above 80% (or any percentage that you specify) then it will be the same message or almost the same message. Then you know its SPAM.

[john278]

Yep, that is what I want. I have tried your script and even though the words don't match it still says "spam".

$text = "nothing about it most that most don't know which ones don and just discovered normally thinks they can make quick buck two well's not";

$words = "games,dinner,whats to be or not to be";

//Convert Keywords to array
$words = explode(',',$words);

//Spam Bool
$spam = false;

//Check Words
foreach($words as $w)
{

  if(stripos(&$text,$w)!=-1)
  {
    $spam = true;
    break;
  }

}

//Check Bool
if($spam==true){
echo 'spam'; 
} 

Hope you can fix it.

Thanks!

[matsko]

Yeah the stripos returns false if nothing is found. So replace...

if(stripos($text,$w)!=-1) 

with

if(is_numeric($text,$w)) 

HOWEVER, read my first post about what I said after the code...

[john278]

It is now saying that I have the wrong parameter count for is_numeric(), which I think is correct as I think you can only use 1.

Also, I want to use this script to stop articles going into my site that are about adult related stuff so that why I mainly want to use keywords. If I have to also use urls, then I will as it should also accept urls in the string that will match the urls in any of the articles description.

[matsko]

Oh sorry my mistake...

if(is_numeric(stripos($text,$w))) 

Yeah I was mearly refering to the text spam filter system. However keywords like G.A.M.E.S. and so on will still be considered not spam...

[UFTimmy]

Oh sorry my mistake...

if(is_numeric(stripos($text,$w))) 

Yeah I was mearly refering to the text spam filter system. However keywords like G.A.M.E.S. and so on will still be considered not spam...

This solution will work, but it's not the optimal solution.

stripos will return boolean false when it doesn't find it.

So you should check for:

stripos( ... ) !== false 

[john278]

Thanks, that works just fine.

[noddy]

Another way that may have worked and been easier possibly would be to have the words in an array. Then use the in_array function to then check if if the words you are looking for exist. However say you were searching through a lot of text this might not be very fast as it would check each submitted word against the list of banned words. This forum has a feature that does this task.