Please fellow

Lalasticlala, dragnet please help me

POS tagging falls under NLP (Natural Language Processing).

https://en.wikipedia.org/wiki/Part-of-speech_tagging

There are plenty of freely available NLP toolkits that will do what you need - just do a simple Google search (or simply look at the end of the Wikipedia page I've linked to).

omohayek:
POS tagging falls under NLP (Natural Language Processing).

https://en.wikipedia.org/wiki/Part-of-speech_tagging

There are plenty of freely available NLP toolkits that will do what you need - just do a simple Google search (or simply look at the end of the Wikipedia page I've linked to).

Thanks

Suen, Lalasticlala, dragnet, puskin please help a Nairalander

Dude! Your Igbo/Yoruba must be spot on. The thing that you asked for is not hard if you know the languages (igbo/yoruba). Start by researching on NLP and POST for english (Part Of Speech Tagging). Search Github for an open source project and just redo it in Igbo/Yoruba.

NOTE: No one here will help you with your project. No one here is gonna get honoured with a BSC from your college. So, if I were you, i'd better start searching. Pls, don't go to other programming forums to ask this. You'll just get burned.

getfupre:
Dude! Your Igbo/Yoruba must be spot on. The thing that you asked for is not hard if you know the languages (igbo/yoruba). Start by researching on NLP and POST for english (Part Of Speech Tagging). Search Github for an open source project and just redo it in Igbo/Yoruba.

NOTE: No one here will help you with your project. No one here is gonna get honoured with a BSC from your college. So, if I were you, i'd better start searching. Pls, don't go to other programming forums to ask this. You'll just get burned.

Thanks for your contribution bro..... I didn't say they should help me with my project.
I just wanted a base and guide and it's allowed in all part of the world and that's why there is space for "Work cited or Reference citation"

And for the searching am already doing that cos that's why am here.. Thanks man

The way I see it, you have two options to implement POST for Igbo and Yoruba.

1. Rule-based algorithm. This option basically means that you will hand-code the syntax and grammar rules that will be used by the algorithm for tagging. Naturally, this requires that you have a good grasp of Igbo and Yoruba... or, at the least, are able to find someone (or some people) who do[es] and is[are] willing to help you. Considering that this is for your final year project, the rule-based algorithm will probably be the most direct route for you. You could probably get away with tagging just the major parts of speech (nouns, pronouns, verbs, auxiliaries, infinitives, etc). I doubt your instructor would be expecting a commercial grade product from you.

2. Machine learning algorithm. This option mostly removes the requirement of you knowing the syntax and grammar rules for Igbo and Yoruba. If you aim to impress your instructor, then this is probably the route you should consider. However, this route also has its own requirements. For one, you will need a large amount of text data. Most machine learning algorithms work on statistical inferences. To make your tagging as accurate as possible, you will have to feed the program with as much text data as possible. Otherwise, the results may be lacking. You still will need to have some knowledge of Igbo and Yoruba so as to teach the program what terms are nouns, pronouns, etc. It also goes without saying that this option requires you have some statistical knowledge in order to build your algorithm (unless you can find an open source algorithm that will work for you).

Regardless of whichever route you take, you will need to know a few things.

1. Regular expressions for splitting the text up appropriately.
2. Some level of understanding of the language (or you could outsource this -- get someone who knows to help you)
3. Figure out the path you want to take and then write your algorithm.
-- 3a. If it is hard coding the rules, then you need to know those rules and decide on how to structure your decision tree
-- 3b. If it is machine learning, then you need an understanding of statistics and a large amount of textual data for the program to process
4. I would recommend a scripting language for this.

ChinenyeN:
The way I see it, you have two options to implement POST for Igbo and Yoruba.

1. Rule-based algorithm. This option basically means that you will hand-code the syntax and grammar rules that will be used by the algorithm for tagging. Naturally, this requires that you have a good grasp of Igbo and Yoruba... or, at the least, are able to find someone (or some people) who do[es] and is[are] willing to help you. Considering that this is for your final year project, the rule-based algorithm will probably be the most direct route for you. You could probably get away with tagging just the major parts of speech (nouns, pronouns, verbs, auxiliaries, infinitives, etc). I doubt your instructor would be expecting a commercial grade product from you.

2. Machine learning algorithm. This option mostly removes the requirement of you knowing the syntax and grammar rules for Igbo and Yoruba. If you aim to impress your instructor, then this is probably the route you should consider. However, this route also has its own requirements. For one, you will need a large amount of text data. Most machine learning algorithms work on statistical inferences. To make your tagging as accurate as possible, you will have to feed the program with as much text data as possible. Otherwise, the results may be lacking. You still will need to have some knowledge of Igbo and Yoruba so as to teach the program what terms are nouns, pronouns, etc. It also goes without saying that this option requires you have some statistical knowledge in order to build your algorithm (unless you can find an open source algorithm that will work for you).

Regardless of whichever route you take, you will need to know a few things.

1. Regular expressions for splitting the text up appropriately.
2. Some level of understanding of the language (or you could outsource this -- get someone who knows to help you)
3. Figure out the path you want to take and then write your algorithm.
-- 3a. If it is hard coding the rules, then you need to know those rules and decide on how to structure your decision tree
-- 3b. If it is machine learning, then you need an understanding of statistics and a large amount of textual data for the program to process
4. I would recommend a scripting language for this.

Thank you so much bro... You just nailed it. Please I really need you so that you can explain some other things for me cos from your explanations, you know much already

And for my supervisor/instructor he is really a hard nut to crack. Please contact my signature since you didn't drop your contact.

Is your course computational linguistics?

enieme:
Is your course computational linguistics?

is Computer Science o my brother

Nsonaso, I really don't know that much. My explanation was general, because that is really where my concrete knowledge ends. Beyond the general explanation, it becomes a question of implementation, and implementation is a think-as-you-go sort of thing. If you still want to ask questions, I don't mind trying to answer. I don't mind thinking it through along with you. Ultimately however, it is your final project. You can't expect myself or anyone else to hold your hand through this.

Nsonaso:
is Computer Science o my brother

oh okay. But that's nice

ChinenyeN:
Nsonaso, I really don't know that much. My explanation was general, because that is really where my concrete knowledge ends. Beyond the general explanation, it becomes a question of implementation, and implementation is a think-as-you-go sort of thing. If you still want to ask questions, I don't mind trying to answer. I don't mind thinking it through along with you. Ultimately however, it is your final project. You can't expect myself or anyone else to hold your hand through this.

Assuming I pick the method rule based algorithm, how do I hand-code the syntax and grammar?

Assuming you opt to go the rule-based route, then what you will need to do is learn the syntax and grammar rules for Igbo and Yoruba (or get someone to help you with this). The essential question you need to answer is: How is a sentence formed? In other words, you need to be able to identify where the parts of speech go in a sentence. You then develop your algorithm based on the grammar and syntax rules of the language. The algorithm will more or less follow your thought process for identifying which part of speech a term belongs to, based on the previously identified grammar and syntax rules.

ChinenyeN:
Assuming you opt to go the rule-based route, then what you will need to do is learn the syntax and grammar rules for Igbo and Yoruba (or get someone to help you with this). The essential question you need to answer is: How is a sentence formed? In other words, you need to be able to identify where the parts of speech go in a sentence. You then develop your algorithm based on the grammar and syntax rules of the language. The algorithm will more or less follow your thought process for identifying which part of speech a term belongs to, based on the previously identified grammar and syntax rules.

For the grammar rule and sentence, I think Igbo dictionary is gonna help.... I want to know how to hand code it.

Please I will really like if we chat on whatsapp or something i.e if you don't mind

By hand-code, I simply mean writing a program with the explicit tagging rules based on the language's grammar and syntax rules. That's all.

ChinenyeN:
By hand-code, I simply mean writing a program with the explicit tagging rules based on the language's grammar and syntax rules. That's all.

OK... So which language do you suggest I use in the hand-code?

Practically any scripting language will do, because (if I'm not mistaken) all of them support regular expressions. So, the choice of which language really depends on you and your level of comfort with said language.

I did something on NLP recently, it was to extract multiple choice questions from a text material.

You could use Stanford NLP tools to build your tagger, I have seen taggers in Chinese and Spanish. Google Stanford NLP tools and try and look for materials on that. I would be interested in your research.

Things you would need
1. A large data set of yoruba/Igbo to create your corpus.
2. Stanford NLP tools to create the tagger.

Best of luck

With my understanding on Human language Processing, what you need to research on or know are some of the following

Firstly, you need to understand the domain of the source language (eg English language) & target language (eg Igbo language).

Secondly, know the structure of grammar of both language. For simple English expression, it uses SVO(subject very Objects). I don't know what Igbo language used but Yoruba language use SVO for simple sentence also.

Thirdly, you design a production rules or re-write rules for both languages & generate a parse tree for them.

After that, you can now decided which approach to use for implementing the Machine Translation.

There are three general approach :
1. statistical based approach which sometimes called data driving
2. Rule based approach which based on context free grammar
3. Hybrid (both 1 & 2 combined).

Note
search for Dr eludiora & Dr Odejobi online for the book on Human language Processing.

Also, if you know about programming, I can send some code I wrote for my assignments using python to translate English sentence to Yoruba sentences.

I used Human language Processing to substitute for Natural language Processing bases on my understanding on both but you can't get anything online about human language Processing. If you need more explanation on that, I can.

Best of luck

Based on some comments above, using python language for your implementation is good expecially when dealing with this kind of project.

I have used python & java to implement something like this & I know the line of codes I wrote for java compared to python.

You can Google for best language online for human language Processing. Python has a lot of support for that.

Note: I don't like using python but that what my lecturer specified to use for him.

Heyzed:
With my understanding on Human language Processing, what you need to research on or know are some of the following

Firstly, you need to understand the domain of the source language (eg English language) & target language (eg Igbo language).

Secondly, know the structure of grammar of both language. For simple English expression, it uses SVO(subject very Objects). I don't know what Igbo language used but Yoruba language use SVO for simple sentence also.

Thirdly, you design a production rules or re-write rules for both languages & generate a parse tree for them.

After that, you can now decided which approach to use for implementing the Machine Translation.

There are three general approach :
1. statistical based approach which sometimes called data driving
2. Rule based approach which based on context free grammar
3. Hybrid (both 1 & 2 combined).

Note
search for Dr eludiora & Dr Odejobi online for the book on Human language Processing.

Also, if you know about programming, I can send some code I wrote for my assignments using python to translate English sentence to Yoruba sentences.

I used Human language Processing to substitute for Natural language Processing bases on my understanding on both but you can't get anything online about human language Processing. If you need more explanation on that, I can.

Best of luck

Thanks man, I really appreciate your effort. Please can I have your contact? You can send it to mine on my signature

Fadman4real:
I did something on NLP recently, it was to extract multiple choice questions from a text material.

You could use Stanford NLP tools to build your tagger, I have seen taggers in Chinese and Spanish. Google Stanford NLP tools and try and look for materials on that. I would be interested in your research.

Things you would need
1. A large data set of yoruba/Igbo to create your corpus.
2. Stanford NLP tools to create the tagger.

Best of luck

Thanks man