Page 125 - Special Topic Session (STS) - Volume 2

STS466 Sasongko Y.
methodology in order to perform various processing of the natural language
data in the media content with objective to establish some context of the data.
The NLP objective is to extract objects in the content and classify them into
grammatical role. Starting from identifying the object as NOUN or VERB, until
the process to identify the function of the sentences, such main-clause and
sub-clause. We also perform the NLP for advanced computational linguistics
by classify the name of person, name of location, people statements,
summarization, etc.

Hybrid approach between machine learning and dictionary based are used
as the framework of NLP model. We leverage on the BDA to perform in-
memory processing in establishing the NLP model using hundreds of
thousand data training. We manage to establish NLP model for English, Malay
and Indonesian language, to be used in media analysis.

3.1.3 Issue Analysis
After the unstructured data has been processed using the NLP framework,
statistical analysis is performed to develop issue analysis based on the context
of media characteristics. Combining both liner and non-linier analysis, we
perform various data story-telling such as:

 Topic Management: Clustering the contents into topic of interests
 Analysis Dashboard: Providing summary of analysis for certain topic
 Influencer Analysis: Detecting list of people that likely having the ability
to influence perception
 Sentiment Analysis: Providing perception sentiment on positive/negative
 Ontology Analysis: Generating relationship among people based on
specific context and problem definition
 Topic Similarity: Establishing the similarity of one issue with others and
develop the similarity behavior and characteristics
 Social Network Analysis: Developing the social network analysis based
on conversation in social media

114 | I S I W S C 2 0 1 9