Reference no: EM132785770

SIT772 Database and Information Retrieval - Deakin University

Assessment - Information Retrieval Techniques Problem Solving Task

Learning Outcome 1: Demonstrate data retrieval skills in the context of a data processing system.

Purpose
This task evaluates the student's technical skills in the management of unstructured data, with potential usage in real applications. This assessment supports student understandings of the techniques related to unstructured data management and data processing

Question 1 (Index Construction):

Suppose you have joined a search engine development team to design a search algorithm based on both the Vector model and the Boolean model.

You have collected the following documents (unstructured) and plan to apply an index technique to convert them into an inverted index.

Doc 1: Information retrieval is the science of searching for information in a document, searching for documents themselves, and also searching for the metadata.

Doc 2:As a critical aspect of Web search engines, the field of Information Retrieval includes almost any type of unstructured or semi-structured data.

Doc 3:The information retrieval system is also made up of two components: the indexing system and the query system.

To answer the below questions, you have to provide the detailed procedures step by step. You need to remove all stop words and punctuation before the process of creating the inverted index. After that, please complete the following steps:

Question 1.1:
Create a merged inverted list including the within-document frequencies for each term.

Question 1.2:
Use the index created as above to create a dictionary and the related posting file.

Question 1.3:
Please design three Boolean queries, (for example, web AND search) and list the relevant documents for each query. Each query must contain at least two keywords while no one keyword appears in one document only.

Question 1.4:
Please use the Vector model to query on the inverted index, and compare the result with the Boolean model. (Hint: you can use cosine similarity and set a similarity threshold).

Question 2 (IR Evaluation):

In this question, you are required to evaluate the performance of different search engines.
• First, please select two of the three search engines you are familiar,

• Second, you are required to constrain your query to www.reuters.com by specifying: e.g., Given a keyword query "high-tech global", you need to write into the search box with "high-tech global site:www.reuters.com" at your selected search engine website. Below is the example provided for you to use google, shown in Figure 2 and Figure 3. [Penalty is applied if your results do not follow the instruction.]

• Third, please choose one target from the following list, and design two queries to search in both search engines. So both query 1 and query 2 have to be tested in both search engines.
• Target 1: Online tutors have more incomes in worldwide education marketing.
• Target 2: Online education expanding, awaits innovation.
• Target 3: Innovations in online learning for regional people.

• Finally, select the first 20 results in both search engines, if they return the target news, then mark them as relevant documents, otherwise, they are irrelevant. Note: assume there are 12 relevant documents in total (retrieved and not-retrieved). If you search more than 12 relevant documents, you can simply regard some to be irrelevant for practice.

The following questions are based on your search results.

Question 2.1:
List your target, results and designed search queries (You can use any keywords you think are related to the target news, even if the keywords are not contained the news text). For each result, you can click the link and go to the page, and write down a summary with 200 words for any 3 results if you think this result is relevant. At your report, you are required to provide the URL address and the summary to explain why they are relevant to the queries.

Question 2.2:
Get the precision and recall values for 20 documents for query 1 in search engine 1. Interpolate them to 11 standard recall levels. Then plot them into a chart. Get the precision and recall values for 20 documents for query 1 in search engine 2. Interpolate them to 11 standard recall levels. Then plot them into the same chart.

Question 2.3:
Get the precision and recall values for 20 documents for query 2 in search engine 1. Interpolate them to 11 standard recall levels. Then plot them into a chart. Get the precision and recall values for 20 documents for query 2 in search engine 2. Interpolate them to 11 standard recall levels. Then plot them into the same chart.

Question 2.4:
Now find the average interpolated precision of query 1 and query 2 for search engine 1 and plot it into a chart. So you will have total of 3 interpolated curves in one single chart. Now find the average interpolated precision of query 1 and query 2 for search engine 2 and plot it into the same chart. So, you will have total of 3 interpolated curves in one single chart.

Question 2.5:
Plot the average interpolated values for Search Engine 1 and Search Engine 2 on one single chart, and compare the algorithms in terms of precision and recall. Which search engine do you think is superior? Why?

Question 3 (Innovation Concept Design of Decentralized Web Search Engine):

In this question, you are required to make the great brainstorming for concept design. As we learnt, all the current web search engine companies host the web data in their own web server. The data generators cannot control their own data. In the near future, it might be highly desirable for the worldwide researchers or companies to design a new type of web search engine, denoted as decentralized or distributed web search engine. Lots of start-ups also make investment in such area. The benefit is to achieve data privacy protection, data use transparency, and removing the centralization of data management.

In this challenging task, your design should include several important aspects:

How to maintain the data using index? Must have at least 500 words to describe the design.

Question 3.2. How to answer keyword queries in the new types of data environment? Must have at least 500 words to describe the design.

Question 3.3. How to evaluate such new web search engine system? Must have at least 300 words to describe the design.

Question 3.4. Provide the system structure concept design diagram. Must have a diagram and at least 200 words to explain the diagram.

Attachment:- Assignment_Description.rar