MLIM Information Retrieval Theory and Practice

Annotated Knowledge Graph

In this assignment, you will create an annotated knowledge graph to explore the process of

information retrieval. Particularly You will explore how different information is

interconnected and how knowledge is embedded in a networked fashion. It is a graded task

that provides you an opportunity to explore the topic of interest in a systematic way with the

support of the network theories we have discussed in class.

Ask a Question

You will need to identify one question that can be answered by empirical research. That is,

the question has to be, at least somewhat, answerable through acquiring experiences from the

real world (e.g., observation, measurement, surveying, experimentation). In this assignment,

you do not have to start until you have the best research question. It is often an iterative

process facilitated by searches.

Identify Relevant Academic Publications

You will identify 5-10 relevant academic publications (e.g., journal articles, book chapters,

conference proceedings) about the question you have asked. This is most likely not going to

be a one-shot search and pick the highest ranked 5-10 papers. Please document the process

and iterations of your searches (e.g., search engines, queries, results, evaluation criteria). You

will have to read these papers for later discussion.

Once you have a collection of relevant documents. Write one or two sentences about each

article to describe what it is about. In the meantime, document the basic information of the

documents (e.g., author, title, keywords). You can make some adjustments if needed: For

example, combining some very similar keywords or use the thesauri/descriptors provided by

the bibliographic database.

Network Analyses

There are a few different network analyses you can choose to do once you have your data.

Choose at least one or more from the following list:

(1) Co-occurrence of keywords/words in title/abstract: Understand how topics are

linked and grouped.

(2) Collaborator network: Understand how the authors work with each other. You

may pull more related articles from some authors’ profile.

(3a) Co-citation network: Explore how these publications may be connected through

ConnectedPapers (https://www.connectedpapers.com/).

(3b) Co-citation network: Analyze the reference list and how the knowledge base of

these publications are connected.

MLIM Information Retrieval Theory and Practice

Zhichun Liu, Ph.D.

(4) Multimodal network: Combine the networks mentioned above.

You are encouraged to be creative (e.g., connect the keywords with documents and then

implement tf-idf to assign weight)!

Depending on your preference, you can choose to the following:

(1) Old-fashioned paper-and-pencil or sketch board

(2) Online visualization tool, such as Nocodefunctions (https://nocodefunctions.com/)

or Flourish (https://app.flourish.studio/)

(3) Local network analysis software, such as Gephi (https://gephi.org/) or UCINET

(https://sites.google.com/site/ucinetsoftware/home?pli=1)

(4) Network analysis packages such as igraph in R or networkX in Python

The latter two options will allow you quantitatively analyze the network data, but it is

up to you.

Discussion

In the discussion section, you will reflect on the results of your searches and analyses.

Consider some of the following questions

(1) Was your initial/revised search strategy effective? What does your impression of

reading papers tell you about the effectiveness and what does the analyses tell you

about the effectiveness?

(2) How does your network analysis results compare to your impression of reading

papers?

(3) Are there any interesting insights you might get from your network analyses?

(4) What are some implications of network analysis? Remember, this assignment is

only at the scale of 5-10 papers. Look beyond the scope of it.

Submission Guidelines

The deliverable of this assignment is a written essay that includes at least the four

components (i.e., question, search process and results, network analysis, and discussion). The

written narrative should not exceed 1200 words (excluding tables, appendices, and

references). The written narrative contributes to 15% of your grade. The essay should be

submitted no later than February 18, 11:59 pm.