Scribd
Upload
10 views2 pages

Web Intelligence Assignment Analysis

Uploaded by

aryanraopippal1
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
10 views2 pages

Web Intelligence Assignment Analysis

Uploaded by

aryanraopippal1
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Web Intelligence

Assignment # 1
1. Online Bookstore Analysis: An online bookstore wants to recommend books based on browsing and
purchase history.
a) Explain which web mining technique (content, structure, usage) would be most effective and why.
b) Two users A and B purchased the following books:
• User A: {AI, Machine Learning, Data Mining, Big Data}
• User B: {AI, Machine Learning, Cloud Computing, NLP}
Compute the Jaccard Similarity between User A and User B.
c) Represent each user as a binary vector for these books: {AI, ML, DM, BD, CC, NLP} and calculate Cosine
Similarity.

2. E-commerce Product Scraping: A company wants to analyze competitor products by comparing reviews.
a) Discuss how scraped text data can be processed for analysis.
b) Suppose two products have been reviewed with the following word counts for {Good, Bad, Cheap, Quality}:
• Product A: [3, 1, 0, 2]
• Product B: [2, 0, 1, 3]
Compute the Euclidean Distance Similarity between Product A and B.
c) Interpret the result for product similarity.

3. Social Media Monitoring: A political campaign team wants to compare public opinion on social media
posts.
a) Explain the importance of text preprocessing before similarity analysis.
b) Consider two preprocessed tweets represented as vectors:
• Tweet 1: [1, 2, 0, 1, 0]
• Tweet 2: [0, 1, 1, 1, 1]
Compute the Cosine Similarity between these tweets.

4. Academic Research Portal: A portal ranks research papers based on query similarity.
a) Explain how similarity functions can improve paper ranking.
b) Given a query Q = [1, 0, 1, 0] and two research papers:
• Paper A = [1, 1, 0, 1]
• Paper B = [1, 0, 1, 1]
Compute Cosine Similarity between Q and both papers and decide which is more relevant.

5. Digital Marketing Campaign: A fashion brand runs an ad campaign.


a) Define Impressions, Clicks, Conversion Rate in web analytics.
b) If an ad had 40,000 impressions, 4,000 clicks, and 600 conversions:
• Compute CTR (Click-Through Rate).
• Compute CR (Conversion Rate).

6. Website Redesign Impact: A travel agency redesigned its website to increase engagement.
a) Explain how Benchmarking helps compare old vs. new site performance.
b) If the old site had 15,000 visitors and 9,000 bounced, while the new site had 20,000 visitors and 7,000
bounced:
• Calculate the Bounce Rate before and after redesign.
• Determine the % improvement.

7. Visitor Segmentation: An education platform wants to analyze visitor behavior.


a) Define New, Returning, and Referral Visitors with examples.
b) If New Visitors = 10,000, Returning Visitors = 6,000, and Bounced Visitors = 5,000 out of the total 16,000,
compute the Bounce Rate.

8. E-commerce Funnel Optimization: An electronics e-commerce website observes cart abandonment.


a) Draw a funnel with these values:
• Product Page Visits = 25,000
• Add to Cart = 15,000
• Checkout Initiated = 8,000
• Purchase Completed = 4,000
b) Compute:
• CTR (Clicks from product page to cart).
• Conversion Rate (final purchases out of total visits).
• Bounce Rate if 10,000 visitors left without clicking further after product page.

Common questions

Powered by AI

Bounce rate is the percentage of visitors who leave a site without interacting. For old site: 9,000/15,000 = 60%, for new site: 7,000/20,000 = 35%. % improvement = (60%-35%)/60% = 41.67% improvement .

Bounce Rate is determined by dividing the number of bounced visitors by total visitors. With 5,000 bounced out of 16,000 total visitors, the bounce rate is 5,000/16,000 = 31.25% .

Preprocessing ensures the text is cleaned and formatted, which eliminates noise and standardizes data for effective similarity measurement. It reduces dimensionality, improves modeling performance, and ensures that comparable components are measured accurately .

Jaccard Similarity is the size of the intersection divided by the size of the union of two sets. For User A {AI, Machine Learning, Data Mining, Big Data} and User B {AI, Machine Learning, Cloud Computing, NLP}, the similarity is 2/6 = 0.33 .

Conversion Rate is final purchases divided by total visits. Given 4,000 purchases and 25,000 initial visits, the rate is 4,000/25,000 = 16%. This reflects the percentage of visitors completing their transactions initiating at the product page .

Cosine Similarity is calculated using the dot product of vectors divided by the product of their magnitudes. Given vectors {3, 1, 0, 2} for Product A and {2, 0, 1, 3} for Product B, their Cosine Similarity indicates the angle between them, showing their relative orientation. This aspect can signal the direction of commonality and differences .

Similarity functions quantify how closely a paper matches a query. By ranking based on these scores, the system ensures the most relevant papers appear higher in results, improving user satisfaction and retrieval efficiency .

Usage mining is most effective for recommending books as it analyzes the navigation and purchase patterns of users, allowing identification of trends and preferences in customer behavior .

Benchmarking evaluates new performance metrics against old ones to identify improvements or regressions. It provides a structured comparison framework, highlighting areas of success and necessitating action points for further enhancement .

CTR is calculated as the number of clicks divided by the number of impressions. For an ad with 40,000 impressions and 4,000 clicks, CTR = 4,000 / 40,000 = 0.1 or 10% .

You might also like