(Web Mining) Assignment 2

Date: 27/11/2019
Report 2
Grade: M1
Name: Roberto Espinoza
Chamorro
Student ID: 6930-31-1295
1. What’s fake news?

Fake news is commonly defined as a form of stories or news that are intentionally and
verifiably false, that gain traction in the culture through via traditional news media or online
social media.1
They are written and published usually with the intent to mislead in order to damage the
reputation of an entity or gain a financial or political advantage using sensationalist,
dishonest, or fabricated headlines.2
Fake news classifications can be mapped out according to two dimensions, from high to low:
1) Facticity, which is the degree to which fake news rely on facts.
2) Author’s immediate intention to deceive.
Classifications can be grouped as follows:
1-High, 2-High: Manipulation (Adjustments, Additions or Removals; includes visual news),
Native Advertising (biased news meshed with advertisement, clickbait) and Propaganda
(news stories created by political entities).
1-High, 2-Low: News Satire (use of humor or exaggeration to present news updates).
1-Low, 2-High: Fabrication (no factual basis, strictly designed to deceive)
1-Low, 2-Low: News Parody (use of non-factual information to present humor)3
2. Why detect and classify fake news is important?

Because fake news is spreading in an increasing rate with the also growing changes in
technology, and the more they spread, the bigger the impact they may have, specially if they
have a harmful objective. Fake news not only spread mis-information, dis-information and
mal-information; they can have real consequences. It is important to be able to detect fake
news since the “information” obtained through them impact in the decisions of the people.
The harder it gets to detect fake news, the harder it will be for people to trust any kind of
1
Tandoc, Edson & Lim, Zheng & Ling, Rich. (2017). Defining “Fake News”: A typology of scholarly
definitions. Digital Journalism. 1-17. 10.1080/21670811.2017.1360143.
2
Hunt, Elle (December 17, 2016). "What is fake news? How to spot it and what you can do to stop it". The
Guardian. Retrieved November 20, 2019.
3
Tandoc, Edson & Lim, Zheng & Ling, Rich. (2017). Defining “Fake News”: A typology of scholarly
definitions. Digital Journalism. 1-17. 10.1080/21670811.2017.1360143.
Date: 27/11/2019
information available. Or on the other extreme, they might believe any kind of news if they
make it way into a trustable source (something that is becoming more usual with the
widespread reach of social networks). Such was the case of fake news spread all along
Facebook by as many as 60 million bots that were infesting the social network. They had
influence because of the substantial posts of political content and were also used to attempt to
influence in the 2017 French election.4 On that note, according to a global survey done by
CIGI-Ipsos to more that 25000 internet users over two dozen countries, 86% of them said
they had fallen for fake news at least once, with 44% saying they sometimes or frequently
did. Only 14% said they had “never” been duped by fake news. 5 Considering this, being able
to classify fake news will help to alleviate or even stop their possible impact. Still, for
identification and classification it is required to understand the common traits as to be able to
recognize them as fake news.
3. How to detect and classify fake news?

Will there is still a lot of work in progress in the field of fake news detection, there are some
technologies defined as state-of-the-art ones that are a key component of fake news detection.
There are two major categories defined for this: linguistic cue approaches (with machine
learning) and network analysis approaches.6
1) Linguistic approaches: Goal is to look for instances of predictive deception cues
(language anomalies in frequencies and patterns) in the content of the message. Some of
the options currently used are:
- Data Representation: Also known as the “bag of words approach” where each word
represents a unit, and individual words of certain frequencies are aggregated and
analyzed. It is dependent on the language and doesn’t consider context information.
- Deep Syntax Analysis: Implemented through Probability Context Free Grammars
(PCFG), which is used to distinguish rule categories (lexicalized, unlexicalized,
parent nodes, etc.)
- Semantic Analysis: Analyzes the degree of compatibility a personal experience as
compared to a content profile derived from a set of analogous data, as to find
contradictions or omissions of facts (key words consists of attribute: descriptor pair).
- Classifiers: Prediction of instances of future deception on the basis of numeric
clustering and distances.
2) Network Analysis approaches: Analysis of network properties and behavior. Some of
them are:
4
E. Ferrara, First Monday 22, 2017 (2017).
5
CIGI-Ipsos. 2019. “2019 CIGI-Ipsos Global Survey on Internet Security and Trust.”
www.cigionline.org/internet-survey-2019.
6
Conroy, N. J., Rubin, V. L. and Chen, Y. (2015), Automatic deception detection: Methods for finding fake
news. Proc. Assoc. Info. Sci. Tech., 52: 1-4. doi:10.1002/pra2.2015.145052010082
Date: 27/11/2019
- Linked data: Analyze the distance of the information node to an existing knowledge
network or publicly available structured data. The closer the node, the higher the
likelihood of the statement to be true.
- Social Network Behavior: Use of a network-based text analysis called Centering
Resonance Analysis (CRA), which represents the content of large sets of texts by
identifying the most important words that link other words in the network
4. Are the methods of opinion mining and sentiment analysis useful for this task? Why?
Considering that there is already semantic analysis and use of classifiers approaches, both
methods reviewed in class could be useful in the task of detecting fake news, if they aren’t
being used already. While syntactic patterns may be used to find anomalies, in the same way,
there are subtle differences of sentiment that can be present in fake news, since there is
pattern in the exaggeration of the sentiment (both positive and negative) that deceivers are
trying to convey on the receivers, which may be even more noticeable in the case of political
topics.

(Web Mining) Assignment 2

Uploaded by

Copyright:

Available Formats

(Web Mining) Assignment 2

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Web Mining) Assignment 2

Uploaded by

Copyright:

Available Formats

Date: 27/11/2019

1. What’s fake news?

2. Why detect and classify fake news is important?

3. How to detect and classify fake news?

You might also like