Last Change Aug 31, 2016. THIS PAGE IS UNDER CONSTRUCTION. Document hand1.pdf contains the LATEST information about this course. Until contents are finalized this web-page might reflect material of another course or OF A PREVIOUS SEMESTER OFFERING. PAGE UNDER CONSTRUCTION. MATERIAL CHANGES. It will be finalized on Sep 6, 2016 @ 9:30am.

An introductory course on web-searching. Information vs data retrieval. The architecture of a search engine. Web crawling. Processing text (tokenization, stemming, stopwords, link analysis and markup). Ranking algorithms based on indexes and links (eg. Kleinberg"s HITS, Google"s PAGERANK). Retrieval Models. Search engine evaluation. Case studies (e.g. Google cluster architecture).

INSTRUCTOR:	Alex Gerbessiotis	E-MAIL:	alexg+cs345@njit.edu
OFFICE:	GITC 4213, 4th floor	TEL:	(973)-596-3244
OFFICE HOURS:	Mon 4:00-5:30pm and Tue 4:30- 5:30pm
OFFICE HOURS:	By appointment Mon/Tue/Wed
CLASS HOURS:	Tue 13:00-15:55 , CKB207

Web Page: http://www.cs.njit.edu/~alexg/courses/cs345/index.html

The following also works: http://web.njit.edu/~alexg/courses/cs345/index.html

1.2 Course Administration

Prerequisites

CS 280 and one of CS 241/CS 252. Last 4 digits of your NJIT id.

Textbook

Search Engines: Information Retrieval in Practice by B. Croft et al., Addison-Wesley, ISBN-10: 0136072240, 2010.

CourseWork:

2 exams (including the final); Homeworks

Grading:

1000 points = Exam1(333) + Exam2(333) + Points-of-HW(334). HW1-HW4 are ordinary homeworks, HW5 is a paper presentation, and HW6 is a programming project. HW5, the paper presentation requires a 20-minute reservation slot to be booked in advance, a one-page summary advance submission (see homework for details) and presentation. For HW6 ONLY, a maximum of three students can work together and each one would collect the assigned graded points. Each HW is at least 66 points.

Exams

All exams are open-textbook only. You may bring a hard-copy of the textbook but you are not allowed to borrow one during the exam or bring in class other material. Exam1 is on Tue Oct 25, 105mins. Exam2 is on Final Week, 120mins on a date to be announced by the Registrar.

ExamConflicts

Per University regulations.

Due Dates

Homeworks HW1-HW4 are by email submission before or on 23:59:59 of the due date. They can be an attachment that is a SINGLE .txt file, or SINGLE Word (.doc, .docx) file, or even a SINGLE PDF file. We acknowledge email submissions promptly. It's up to you to properly form and submit an email. Use an NJIT email address and include a Subject line as specified in Handout 0. 11 pts deducted from grade at deadline plus 1 minute, 22 pts every 24hrs thereafter.

Topics

Tentatitive list of topics

• T1 : WebSearching : Introduction
  T2 : Fundamentals of Information Retrieval. 
  T3 : The retrieval process: Crawlers and crawling.
  T4 : Search Engine Architecture, Duplicate Handling
  T5 : Document Processing: Parsing and Tokenization , 
  T6 : Document Processing: Indexing
  T7 : Modeling retrieval and ranking
  T8 : Queries, Query processing, and Interfaces
  T9 : Search engine evaluation
  T10: Classification and categorization
  T11: Google MAPREDUCE model
  T12: Case Studies: GFS
  T13: Other Topics: Social Search
  

A1,B1,F1

Learn the fundamentals of Web searching and be able to communicate succintly in writing and/or orally concepts related to Web searching and the architecture of search engines and Web search engines.

A2,B2,I1

Learn how a Web search engine works and be able to identify and describe the components of its architecture and identify and explain the output results of search engines in the context of web searching.

A3,B3,J1

Learn the requirements and characteristics of web crawling, document, processing and indexing and be able to understand and enumerate and describe the stpes involves in each phase.

A4,B4,C1,I2,J2

Learn how to use fundamental data structures to organize, index and store information for processing web search requests and be able to design a search engine architecture based on input design requirements.

B5,C2,I3,J3

Learn the fundamentals of ranking, and indexing algorithms and be able to understand ranking and indexing algorithms and their limitations; be able to effectively apply ranking algorithms on sample problem instances.

A5,B6,I4,J4

Learn how high performance computing can benefit web searching and be able to apply its methods in the design of a Web search infrastructure.

I1, K1

Learn the fundamentals of web searching and be able to design and implement a desktop-based search engine architecture of varying complexity using available programming language and software tools of your choice.

* First day of classes is the Tuesday (9/6) after Labor Day (9/5) ** Check with the Registrar

Grading

Written work will be graded for conciseness and correctness. Be brief and to the point and write clearly. Programming problems will be graded based on test instances decided by the instructor on an AFS machine (afsconnect1,afsconnect2,osl11). Do not expect partial credit if your code fails to run on all test instances, and you do not provide a bug report.

Grades

Check the marks in written work and report errors promptly. Resolve any issue no later than the Reading Day. For students who submit programming work or have a paper presentation, an email with your grade will be sent back to you. The final grade is decided based on a 0 to 1000 point performance. A 50% or more is C or better, 85-90% or more usually guarantees an A.

Collaboration

Collaboration of any kind is NOT allowed in the in-class exams and homeworks HW1-HW4, HW5. An exception to this rule is HW6 that explicitly allows for collaboration (teams of no more than 3). In such a case collaboration is allowed between members of the team only for the specific homework only. Students who turn in work/answers to questions sourced through the Internet or otherwise, or is product of another person's/student's work, risk severe punishment, as outlined by the University. The work you submit must be the result of your own effort.

Mobile Devices

Mobile phones/devices and/or laptops/notebooks MUST BE SWITCHED OFF (NOT JUST SILENCED) before the class exams. Switch off noisy devices before class.

Email/SPAM

Send email from an NJIT email address. NJIT spam filters or we will filter other email address origins. Use the appropriate subject line as specified in Handbout 0. Include cs345 in the subject line then.

Missing class

If you miss a class and there is no Exam or Homework due it's up to you to make up for lost time.

Missing Exam

If you miss an exam and there is a valid documentation for your absence, such documentation must be presented within 3 working days from the day the reason for the absence is lifted. The maximum accommodation will be the number of missing days to the exam date: it is imperative then that you contact relevant parties even before the 3 working day period has expired if the accommodation period is shorter. You also need to present your case to the Dean of Student Services (DOSS). We will respond after receiving confirmation from DOSS or we can give a make up exam that would only count if DOSS approves and the makeup exam is taken within the accommodation period.

Missing HW

If you are sick (see Missing Exam for the procedure) there is no notion of a make-up homework or delayed submission of a homework other than the penalties specified on page one of this document. Per DOSS and Instructor approvals, a homework grade might get extrapolated from EX1 or EX2.

Programs

Follow submission guidelines for HW6, if you plan to do it.

Presentation

Follow submission guidelines for HW5, if you plan to do it.