"NLP" talk about CRF based on machine learning perspective

talking about CRF based on machine learning perspective

Bai Ningsu

August 3, 2016 08:39:14

"Abstract": the condition with the airport for sequence labeling, data segmentation and other natural language processing, showing a good effect. In Chinese word segmentation, Chinese name recognition and ambiguity resolution and other tasks have been applied. This paper is based on the understanding of the condition with the airport and the application of natural language processing in the process of tagging the sentence recognition sequence. Written mainly from natural language processing, machine learning, statistical learning methods and some of the online data on the introduction of CRF related related, and finally a large number of research and consolidation into the system knowledge. The article is arranged as follows: The first section introduces the basic statistical knowledge related to CRF, the second section introduces the CRF introduction based on the natural language angle, the third section introduces the CRF based on machine learning, the fourth section introduces the relevant knowledge based on the statistical learning angle, and the fifth section introduces the CRF in the depth of statistical learning. ( This article original, reproduced please specify the source : based on the machine learning perspective of CRF. )

Directory

"Natural language Processing: Walking conditions with the Airport series article (i)": Foreplay : Go into the condition with the airport

"Natural language Processing: Walking conditions with Airport series article (ii)": talking about CRF based on natural language processing

"Natural language Processing: Walking conditions with Airport series article (iii)": talking about CRF based on machine learning perspective

"Natural language Processing: Walking conditions with Airport series article (iv)": talking about CRF based on statistical learning

"Natural language Processing: Strolling conditions with the Airport series article (v)": conditional with the airport knowledge expansion

1 conditions with the airport (can be regarded as a given observation value of Marco with the airport)

CRF is a discriminant model of non-direction graph

CRF attempts to model the conditional probabilities of multiple variables after a given observation, specifically, if the order is an observation sequence, and the corresponding marker sequence, then the objective of the CRF is to construct the conditional probability model P (y| X).

Note : The tag variable y is a structural variable, such as in the natural language processing of the sentence labeling task, the observation data as sentences, labeled the corresponding part of speech sequence, with a linear sequence structure, in the parsing, the output marker is a syntax tree, has a tree structure.

The g=<v,e> indicates that the node corresponds to the element one by one in the tag variable y, indicating that the node v corresponds to the tag variable, n (V) represents the knot V's bow tie point, if every variable of Figure g satisfies the Markov nature, i.e.

, then (y,x) constitutes a CRF.

The above formalization in the second chapter has been introduced through the example analysis.

2 Chain condition with airport

As indicated in the above sentence, because of the phenomenon in the application, when the tag sequence modeling, often chain structure ( specific chain structure in front of the introduction)

Similar to Markov random field definition joint probability probability, CRF uses the potential function and the group on the graph structure to define the conditional probability P (y|x) given the observed sequence x, the so-called group is a single tag variable {} and the adjacent tag variable to choose the appropriate potential function, namely the shape:

The conditional probability definition, in which the potential function corresponds to Q, is a normative factor, in practice, Z does not need to obtain an exact value.

In CRF, conditional probabilities are defined as follows by selecting potential functions and introducing feature functions:

The above parameters are explained in detail in the second chapter.

feature functions :

The transfer characteristic function of sentence annotation as an example

When the first observation value is "Love", the relative mark is b,i, and its state characteristic function is as follows:

Indicates that the observed value x is the word "love", the corresponding label is likely to be I

3 references

"1" The beauty of mathematics Wu

"2" machine learning Zhou Zhihua

"3" Statistical natural Language Processing Zongchengqing (second edition)

"4" Statistical learning Method (191---208) Hangyuan li

"5" Network resources

4 Natural language related series articles

"Natural Language Processing":"NLP" revealing Markov model mystery series articles

"Natural Language Processing":the "NLP" Big Data Line, a little: Talk about how much the corpus knows

"Natural Language Processing":"NLP" looks back: Talk about the evaluation of Learning Models series articles

"Natural Language Processing":"NLP" quickly understand what natural language processing is

"Natural Language Processing":"NLP" natural language processing applied in real life

Statement : Regarding this article each chapter, I take the comb main, the smooth bright writing technique. The system reads the related bibliography and the data summary combing, aims at the technology to share, the knowledge precipitates. Thank you for the selfless work of bringing it together in a book. Secondly, my level is limited, the right to use for knowledge accumulation, it is inevitable that the subjective understanding is inappropriate, causing the reader inconvenience, based on this kind of situation, hope readers feedback, easy to correct in time. This article original, reproduced please specify the source : based on the machine learning perspective of CRF.  

"NLP" talk about CRF based on machine learning perspective