H6513 Information Organisation
Types of Analytic Methods
Copyright By PowCoder代写 加微信 powcoder
Dataset preparation
Dataset is usually represented in tabular form with:
rows (records/entities/units of analysis) and
columns (fields/attributes/variables)
CustID Gender Age Race Chocolate Coffee Durian Biscuit
60 Chinese 1 0 0 1
40 Chinese 0 1 1 1
M 30 Chinese 1 1 1 0
F 25 Chinese 0 1 1 0
1. Association analysis
Relations or links between entities / attributes
entity—attribute value
chris—chocolate, dion—sausage, chu_keong—coffee
attribute 1 value—attribute 2 value
chinese—tea, coffee—tea
Association rule mining (directional)
If chris, then chocolate (confidence 80%)
Sequential association rule mining
If bread => butter => jam, then => milk (confidence 80%)
2. Predictive analytics
Models used for prediction or classification (categorization)
E.g.: Gender + Undergrad major + Age
=> MSc programme
Supervised learning is typically used to develop the model:
Linear regression
Logistic regression
Decision tree induction
Neural network modeling
Support vector machine
Random forest
Naïve Bayes
3. Cluster analysis
Cluster similar entities together
E.g. cluster students based on Age, Gender, Undergrad major, Nationality
Cases are clustered based on the principle of maximising the intraclass similarity and minimising the interclass similarity
Unsupervised learning is typically used
4. Time series
Models temporal trends for items whose attribute value changes over time
Usually to predict a future value based on past values, e.g. predicting
stock values
patient’s blood pressure
5. Social network analysis
Model associations among a set of entities using networks and “graph theory”
To identify
network structures & properties
clusters of associated entities
entities occupying certain types of positions in the network
6. Text mining
Analysis of documents and words in them
Association analysis
Predictive analytics, esp. automatic categorization
Cluster analysis
Time series, e.g. trending topics
Social network analysis
Additional methods:
Information extraction
Topic modeling
Social Media Sites as
Sources of Information
Social media
Emphasis on:
interaction between users — conversations or dialogue
user-generated content and
building of online relationships and communities
Types of social media
(Kaplan &Haenlein, 2010)
social networking sites (e.g., Facebook)
blogs and microblogs (e.g., Twitter)—online journaling
discussion forums
collaborative/social bookmarking/tagging sites
consumer review and rating sites
content communities (e.g., YouTube, Flickr)—file and content sharing
3. 非常重要 比如Q & A
4过去很重要 现在不太实用
Other types of social media
collaborative projects (e.g., Wikipedia)—collaborative authoring
virtual game worlds (e.g., World of Warcraft),
virtual social worlds (e.g., Second Life).
online auction sites (e.g., eBay)
voice & video communication services (e.g., skype)
text communication services (e.g., email, instant messaging, whatsapp, line)
This discussion focuses on two types of social media applications where there is a substantial amount of user interaction (i.e. conversation) which is captured as user-contributed content: social networking sites, especially Facebook, and discussion forums.
Challenges faced by researchers
Different types of social media applications support/encourage different types of user behavior & content
The technology and functionality of social media applications are still evolving
Use of social media is evolving, as users figure out what useful things they can do
Challenges (cont.)
Type, quality and quantity of user-contributed content is evolving
Users are gradually discovering what kinds of useful information are available in different types of social media sites
Human sources of information
The most preferred type of information source
they are likely to be consulted first and are most frequently used
Case (2012):
Use of other [information] channels tends to be predicted by the social presence they offer, that is, how much they are perceived as being like a face-to-face conversation with another person, or as Johnson puts it “the extent to which they reveal the presence of other human interactants and can capture the human, feeling side of relationships”
Social media as info source
Social media makes it easier to access human information sources
It also records the interaction and the information for future reference
Human information sources are associated with everyday life information. Similarly social media sites are becoming associated with everyday information
Social media can provide information not found in other Web sources or not retrieved by search engines
Strengths of social media sites as info source
Only humans can provide certain types of information: opinions, advice, recommendations
The information source is personally known to the user (to a greater or lesser extent)
they are trusted sources and have cognitive authority (at least the sources can be evaluated from past experience/record)
Users can provide localized (geographically specific) information, and current or time-sensitive information
Strengths of social media sites (cont.)
Information provided by other users are customized for the requestor
Social contacts can perform intermediary functions of researching, synthesis and packaging of information
Users are able to broadcast a question to a known group of people
Users can obtain emotional and social support
Real nature of information-from human
User behavior of
information integration & knowledge synthesis
Contribution of related, complementary or contrary information, experience and opinion
Reference or linking to external sources of information
Evaluation of information or critical comment
Linking or comparing information in two or more posts
Summarizing a set of posts
Drawing inferences and conclusions from a set of posts.
Types of knowledge synthesis
Kazmer et al. (2014)
Distributed knowledge — no single person has all the relevant pieces of information, but pieces of information from different people are pooled together to provide a more complete understanding.
Undiscovered public knowledge — linking different pieces of information and inferencing new knowledge.
Authoritative knowledge — “co-constructed” by a community of users by examining evidence from a variety of sources.
1.例子：没有一个人拥有关于新冠的全部知识，
Informational + Socio-emotional Support
Information provided on social media tend to be accompanied with socio-emotional support.
Wikgren (2003) found that emotional support and information sharing are closely linked activities:
Communicating health information in virtual communities is, on the one hand, an exchange of facts and a communication of scientific knowledge, but, on the other hand, a communication of meaning and socioemotional support, and perhaps even a “construction of mastery.”
Informational + Socio-emotional Support
Chuang and Yang (2012) found two types of social support in the user postings:
informational support, including facts, advice, information referral, personal stories and opinion
nurturant support, including esteem support, network support and emotional support.
The relative proportion depends on the communication medium — a higher proportion of informational support on discussion forums
Strong + Weak Ties
Social media sites exhibit both strong and weak ties, and weak ties can become stronger over time as users continue to interact online and, in particular, when they decide to meet offline
Granovetter (1973) proposed the theory of weak ties that distant and infrequent relationships are efficient for knowledge sharing and giving access to new information by bridging disconnected groups and individuals, in contrast to strong ties as between family, friends and colleagues.
Social media as a news source
SNS alerts users to breaking news (local, national and international)
On Twitter:
Breaking news is the most popular type of information shared
An important source of news during times of natural and political disasters/crises
for collecting information from disaster scenes and broadcast disaster relief information to the public
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com