Question

İ need a function for a search machine, that this function should find the most-looking like string values to the users input, example: if user writes "Math" this function must find the "mathematics" keyword and match it to the pertinent links and list them, then another function should list them on screen :d

Answer 1

I'm sure this would not be the best way to do it. But what i would do is to make use of AJAX to scan the values as user types in the forum and pass it to a server script (I prefer Python), Then use regular expressions to match it with predefined tags and return those using Javascript and display them back to the browser.

Answer 2

but if theres many datas, i think that that would take fun, so that i think that i can catogarize the names of types under of the users inputs, that will be easier to this function work

i mean that mathematicsz, math problems.. should be under of the catogory maths (keyword) that it will make 2 x faster to search and the functoin will list everything under the catagory

Answer 3

Yeah that would greatly maximize your search efficiency.

Answer 4

errr but i dunno Python :) i know C# and Vb.NET...

Answer 5

an dyou know? i am only 14 years old ^^

Answer 6

That's nice to hear. Learn Python man... I strongly recommend it for its productivity and is really fun. It will take a maximum of 2 days to get accustomed with it and write some highly productive code.

Answer 7

Python is especially good for beginners as it resembles to pseudocode style

Answer 8

ahm but Can i make it to work with .NET? (and i started dto learn programming when I was 12 not a begginner :) coded some 2 d games in C# and XNA :) ) and also know C++ too

Answer 9

but whats the avantage of using Python??

Answer 10

Hmmm. Frankly speaking, I'm not a big fan of M$ and their stuff man. In fact i hate them and i'm inclined to Free/Open Source software. So i am will not be able to Answers your questions on C# or .NET (I'm sorry)

But what i can do is to highlight the advantages of using Python. Python is a very high level programming languages, which unlike C++ handles most of the memory related issues by garbage collection. It has one of the big hacker community crunching out thousands of modules every day. Its officially supported by google and most of the management stuff like build management, source code version control system, even the most hyped google app engine, runs on Python because of High productivity.

It also a lot cleaner than other languages and verbose only up to the required extent (unlike C++ or Java).

Here are some links which you may need to take a look at
http://www.python.org/about/
http://www.linuxjournal.com/article/3882

Answer 11

sorry, had to go out of the building, give me some mins to read it ^^

Answer 12

ok thx and ill learn python as fast as i can (after you understand the logic of the programming its easy to learn languages :D ) but, my webpage will be an .aspx, :(

Answer 13

and i don'T know javascript too :(

Answer 14

and I don'T know much about web progrramming that if you have some usefull links can you write them here?

Answer 15

I strongly recommend you to start web programming by going through the tutorials on HTML, Javascript at http://www.w3schools.com/ . If you are serious on web programming, you will need to know XML, HTML, Javascript, (Java or Python or Perl or Ruby or M$ Stuff like .NET etc..). You clear the basics from the above sites and can then master them by practise :)

Answer 16

Thank you much ^^ do you have an email adress or msn?? i would like to add you mybe ill have some good questions to ask ^^

Answer 17

and err ^^ how can this search engine match catogories with user input? i have some reative ideas

Answer 18

You can contact me at 'srinathsmn <at> gmail.com'. I'm not sure on your question, if the following answer does not suffice, please explain the question with greater verbosity.

When i was in my information retrieval class last semester, i read a concept called stemming, in which words like Mathematics, Maths, Math, Math etc... will be pointing to a particular keyword 'math' or a couple of keywords. based on which the web pages will be classified as. I used the following free book for information retrieval http://nlp.stanford.edu/IR-book/information-retrieval-book.html may be a glance on it may clear your ideas :)

Answer 19

i think that we need an array string[stringnumber(for example 5 in maths)]
it should collate string[stringnumber] with userinput[stringnmber] as strings if there is more then stringnumber/2 strings balanced then it should list the datas in same category with it ^^ and it should also look at uncatogorized keywords too

Answer 20

ty ^^

Answer 21

I'm not sure about your approach for millions of keywords man. I think it would drain your memory. May be you can use some distributed key value stores and then store values like key-value pair (ex) Math = [Maths, Mathematics,Math,....] <which would be a list> or use several layers for categorisation of stem words. I strongly dont recommend the first method as it would be necessary to scan over the entire list of values.

Answer 22

hmm logical

Answer 23

Then tihs will like Catogories>keywords>links YEA?

Answer 24

for example Math>Mathmematics>www.math.com, www.mat4fun.com...
and a list of these links should emerge on screen :D

Answer 25

but if there is milliars of datas and catogories in your database this will harm your engine...
so that namespace>catogories>matematics>links
should be better i think

Answer 26

one namespace can include more then 1 catogory, when it finds the true namepscae it can colalte catogories and then list troe links on screen

Answer 27

Kinda like that, I'm sure that would be more efficient and also learn the index techniques from the link i gave above (information retrieval book). it illustrates some of the best methods to categorise data

Definitely millions of data would drain your engine. Hence you must make use of distributed processing (using frameworks like Hadoop). An open source search engine is known at Nutch is a good project for you to read its architecture on searching. It can be reached at http://nutch.apache.org/ .

Regarding name space , Treat the URL as key which you will eventually end up having millions of links (They will be unique as you are considering the full URL), You then make use of distributed computing infrastructures to crush them for you.

If you are really into search engines. Then you might want to learn about google's infrastructure for storing the indexed data using Bigtable ( http://labs.google.com/papers/bigtable.html)

It would be helpful to gain some practical experience using Redis ( http://redis.io/) a simple key value store in C

Answer 28

thank you much ^^ and will read these as fast as i can ^^

Answer 29

It would be my pleasure to help you in these context as i myself will be working on Data Analysis for the final year project :)

Answer 30

:) which universe you in?

Answer 31

You mean which country? I'm from India

Answer 32

no no the name of the universty you are in (my fault :( )

Answer 33

I'm doing my B.E. in CSE at Thiagarajar College of Engineering, Affiliated to Anna Univ

Answer 34

kk ^^