Textual analysis

I am trying to make a program that can read a question by understanding the type of question, the verb, noun, adjective and tense. This is my summer project... The problem is, I am not really sure here to begin. I am having trouble finding any topics that can help me with this. I read about lexical analysis, and that's the only thing that I can find that seems like it would have any relevance to what i want to do. I just want some one to point me in the right direction. Anyone have any ideas?

 

You'll probably want to make use of some existing toolkit as writing everything from scratch is very time consuming and far less effective (unless your goal is really to learn the core natural language processing concepts).

From the list of such toolkits below I would say that FreeLing is probably the preferred one for you as it seems to have the lowest entry barrier among free C++ choices. If you would reconsider going with Java or Python there are many other toolkits that are easier to learn and even some nice beginner books and tutorials.

http://en.wikipedia.org/wiki/List_of_natural_language_processing_toolkits

 

Thanks Null, had exactly the same questions.

~ DD

 

Thanks, but like you said I am trying to learn as much about c++ and programming concepts as I can. So far what I am trying to do is break down sentences and put each word in a class so I can give words properties (tenses,singual,plural ect...). After that I will put each word item into a linked list for accessibility...

 

Why are you using a linked list? I'm not sure I'm following how you're approaching this, but if you plan on individually analyzing each word against a preset dictionary of sorts it would be better to use a graph structure.

If you're storing it and comparing it in some other way, you'd be better of using a hash table unless you've got a predefined (and small) data size.

Could you perhaps update us with progress you've made so far and some pseudo code?

 

I am still fairly new to C++ and have not done any past programming. I ran a search on graph structures and I do not really understand them. I get hash tables alittle more but I do not understand how address are assigned.

Ok so far I have this.
A class called sentence and it consists of nodes of class Wnodes.
Wnodes consist of a word and a pointer to the next word as it would come in the sentence.
Words have their own class called Word and each word has properties mentioned in my last post.
I still need to find a way to determine and set the properties for each word but that will come after I have everything above set up and working.

What I said above seems a little confusing so I understand if you don't get it. So far, however, from what I understand about C++ a linked list seems like the best way to arrange the words. If you know of a better way please let me know.