In this hands-on lecture, I will discuss about two kinds of parsing and look over them through the called base text minor. For this exercise, I want you to look at Document.Java. Document.Java is different from sentence.Java and token.Java and that it is higher level of representation. So this is a good forward document classification, clustering sentiment classification and so on and so forth. As you see here document is either full text, or or webpage which we call this full content, it's called document. And the document has vectors of the futures which in numeric value. And it has a topic associated with each document. And in the same manner as sentence the java or token the java. It has preprocessed function and preprocessed function process and it has several. Get and step method along with classifications and topic modeling. So we'll talk about this in later in this course. Now, what you need to do is you need to c reate ParsingMain.java. The ParsingMain.java what it will do is it will parse a sentence worl text sentence. Inform of parse tree either constituency parse tree or independent tree. So let's create ParsingMain.java. You right click on main package and select new and select class and then type in ParsingMain. And click on finish and you will have skeleton class of parsing main to java. The parsing main to java as I told you before, they will do two things. One is, it creates constituency parse tree. Second thing is dependence. I'm going to use text on their data, on their purpose. If you text.txt. This is one of text that we created, and we collect it from github.chap, github.com, github.org. Okay? So each data record has your ad on and its description and so on and so forth. So let's go back to passing main data. And you look through, you open up the file with the file, with the scanner and you look through the file until the end of the file. And then you create document. This case document You have text. So text means either set of sentences or abstracts or full text whatever, you collect it. And you click process, after you click process inside document click process function equals stamp for equal LP. Does sentence boundary detection, so sentence boundary detection which produced a set of sentence object. In terms for Stanford CoreNLP we simply take their sentences into our document, our document has a set of sentence. This is how it works. Okay? So for each sentence we create dependency. First constituency parsing. The second dependency person. The sentence as function code get sentence, function code get dependency. The get sentence, let's go there, is simply passing the whole sentence and Get dependencies you create returns dependence tree, which was constructed by CoreNLP. Getparsetree function returns which was generated by Stand for core nlp. Right? So what you need to do is close those two methods, getParseTree and getDependencies and select those two from Sentence object. Save this main function. And after that what you need to do is to simply execute parsing in the Java. Okay, let me explain this result a little bit more. The sentence, for this original sentence you have string representation of tree so which means is collapsed tree and has a one node. That has np np is under the nn is is vp has vvn pp and so on and so forth. So you have this collapsed parse tree which is contingency tree and another one is Dependence tree. So dependence tree, if you recall correctly, in my lecture notes,. I explain that dependence tree has relation between two words, so let's take a closer look at this example. So amod which practice and test these two terms. This number, which means position of that term in given sentence. So this is and the tours has the relations type a mark. X pass thirty pass it's formalized an R. So R formalized has AUXPASS, relation type, and so J is programmer used. Used to locating 17th position programmer is located 15th position of that sentence. So these two terms N sub J relation. There are number of relations here and the relations explain in dependency manual which are provided by Stanford Faculty. So that manual Precisely explain each relation type and describe what it does, and it also provides examples of each relation, dependency relation.
