Chris Pollett > Students > Umaranikar

    Print View

    [Bio]

    [Blog]

    [CS297 Proposal]

    [Google as an ad space seller - PDF ]

    [Credibility of rating on Amazon - PDF]

    [Study of Google's advertising services]

    [configuring Ad Server- Yioop patch]

    [Designing of online auction system]

    [Prototype - Find relevant ads for the search query]

    [CS297 Report - PDF]

    [CS298 Proposal]

    [Online Auction System in Yioop]

    [How System Calculates Min Bid?]

    [How System Updates Bid Amount?]

    [Display Relevant Advertisements]

    [CS298 Presentation - PDF]

    [CS298 Report - PDF]

                          

























CS297 Proposal

An Open Source Ad Server

Pushkar Umaranikar(pushkarumaranikar1@gmail.com)

Advisor: Dr. Chris Pollett

Description:

One prominent revenue(over 80-90%) source for search engines such as Google, Yahoo, Bing, Facebook is display advertisements. Sophisticated algorithms are used to compute the relevance of these ads to a given users query. Using Machine Learning technique, search engines render the most relevant ads to the search query which are very popular and effective.

In addition, pricing of ads to the advertiser is done using sophisticated auctioning systems. A modified version of Dutch auctioning is used, in which the concepts like Cost Per Click are used.

Yioop is an open source search engine software developed by Dr.pollett. It can easily be extended. The aim of this project is to extend Yioop with an open source advertising platform.

Schedule:

Week 1: (Aug 26-Aug 31)Project Proposal.
Week 2: (Sep 1-Sep 7)Understanding of Yioop.
Week 3,4: (Sep 8-Sep 21)Understanding of how online search advertising works.
Week 5: (Sep 22-Sep 28)Deliverable #1:Study of Google AdSense and Google AdWords.
Week 6: (Sep 29-Oct 5)Understanding Auction concept for displaying advertisements.
Week 7,8: (Oct 6-Oct 19)Deliverable #2:Implementation of configuring advertisement server in Yioop
Week 9: (Oct 20-Oct 26)Understanding of bag of words model.Explore online auctioning system
Week 10: (Oct 27-Nov 2)Deliverable #3:Design online auctioning System
Week 11: (Nov 3-Nov 9)Understanding of machine learning model that can be used to exploit the search results to choose most relevant ads.
Week 12,13: (Nov 10-Nov 23)Deliverable #4:POC to show content based ads based on users query
Week 14,15: (Nov 24-Dec 7)CS297 report.

Deliverables:

The full project will be done when CS298 is completed. The following will be done by the end of CS297:

1. Project proposal.

2. Study of how online search advertising works (Sponsored Search). Try Google AdSense and Yahoo AdChoices with Yioop.

3. Build an application to accept information of advertisers.

4. Build a bag of word model to display advertisements for the given user's query.

5. Build a machine learning application to improve most relevant ads based on search results.

6. CS297 report.

References:

1. How Google AdWords work and auctioning concept.
    http://www.wordstream.com/articles/what-is-google-adwords

2. How to predict the Ads having higher probability of clicks:
    https://static.googleusercontent.com/media/research.google.com/en/us/pubs/archive/41159.pdf

3. Search Advertising using Web Relevance Feedback.
    http://gabrilovich.com/publications/papers/Broder2008SAW.pdf

4. How Does Google Sell Ad Space?
    http://books.google.com/books?id=N5DJJXoLPDQC&printsec=frontcover#v=onepage&q&f=false