[Deep Learning Versus Machine Learning] What is the wordnet? Better than RNN Deep Learning?

Machine Learning versus Deep Learning¶

• Hello. I'm DVM who is talking about Deep Learning versus Machine Learning
• Today, I'm going to talk about Machine Learning about Natural Language Processing

In [1]:

from IPython.display import Image
Image('wordnet.png',height='400',width='400')

Out[1]:

• Most of Nautral Language Processing is made by RNN Deep Learning
• Because It is very powerful and accurate than normal machine learning
• But,today will talk about machine learning which name is wordnet.
• The advantage of word net is faster and more simple than deep learning.
• So many people use word net find the keyword

What's the natural language processing

• Natural language processing make computer understand the word
• It have 3 types such as thesaurus, static method and estimated method.
• Today i will talk about Thesaurus.

Let's download NLTK library¶

In [2]:

!pip install nltk

Requirement already satisfied: nltk in c:\users\ssdwin10pc\anaconda3\lib\site-packages (3.4.5)
Requirement already satisfied: six in c:\users\ssdwin10pc\anaconda3\lib\site-packages (from nltk) (1.12.0)

In [3]:

from nltk.corpus import wordnet

Let's Find synonym using wordnet.¶

There are 4 synonym of "AI"¶

In [4]:

wordnet.synsets('AI')

Out[4]:

[Synset('army_intelligence.n.01'),
 Synset('artificial_intelligence.n.01'),
 Synset('three-toed_sloth.n.01'),
 Synset('artificial_insemination.n.01')]

The word I want to talk about is "artificial intelligence"¶

In [5]:

_AI = wordnet.synset('artificial_intelligence.n.01')

Let's find the definition of "AI"¶

In [6]:

_AI.definition()

Out[6]:

'the branch of computer science that deal with writing computer programs that can solve problems creatively'

Let's see what is in AI Group!¶

In [7]:

_AI.lemma_names()

Out[7]:

['artificial_intelligence', 'AI']

Let's find parents and children relation ships¶

In [8]:

_AI.hypernym_paths()[0]

Out[8]:

[Synset('entity.n.01'),
 Synset('abstraction.n.06'),
 Synset('psychological_feature.n.01'),
 Synset('cognition.n.01'),
 Synset('content.n.05'),
 Synset('knowledge_domain.n.01'),
 Synset('discipline.n.01'),
 Synset('engineering.n.02'),
 Synset('computer_science.n.01'),
 Synset('artificial_intelligence.n.01')]

Let's calculate word similarity¶

In [9]:

_CS = wordnet.synset('computer_science.n.01')
_KD = wordnet.synset('knowledge_domain.n.01')
_EG = wordnet.synset('engineering.n.02')

In [10]:

_AI.path_similarity(_CS)

Out[10]:

0.5

In [11]:

_AI.path_similarity(_KD)

Out[11]:

0.2

In [12]:

_AI.path_similarity(_EG)

Out[12]:

0.3333333333333333

So computer science is best keyword¶