Text Analytics:
95-865 (K)
 
CMU logo
 

This page is for Section K (Adelaide).
The page for Section A (Pittsburgh) is here.

Description: Many organizations need to analyze large amounts of text to discover useful information. For example, a company may want to monitor how the public discusses its products in social media, or a forensics team may need to discover the contents of disk drives seized by law enforcement. This course provides students with an understanding of common and emerging methods of organizing, summarizing, and analyzing large collections of unstructured and lightly-structured text ('text analytics'). The focus is on algorithms and techniques, however the course also provides an introduction to open-source software tools
 
This is a 6 unit course. It is offered during Mini-2 and Mini-4.
Learning Objectives: By the end of the course, students are expected to have developed the following skills. Skills are assessed by the homework assignments and the final exam.
  • Recall and discuss common methods of conducting exploratory and predictive analysis of text information;
  • Use search engines and common open-source software to perform common methods of exploratory and predictive analysis; and
  • Apply text analysis techniques discussed in class to solve problems faced by governments and companies;
Prerequisites: None
Time & Location: Mini A4, Fr 3:00-5:40pm, TOR Classroom 3
Instructor: Jamie Callan
Teaching Assistant: Himani Gupta (himanig@andrew)
Office hours: TBD
Office hours are held using a Google hangout. You will need to have a Google account and may need to install a browswer plug-in. If you are unable to use Google hangouts, contact Charles by email to make other meeting arrangements.
Discussion Forum: A discussion forum is provided for students to ask questions, answer questions, and discuss class-related topics. You will need a Piazza account to use the discussion forum. Please provide a CMU email address when you join the 95-865 discussion (you can use other email addresses, too). We will periodically remove students that do not have CMU email addresses.
Instructional Materials: Some lectures have assigned readings from Introduction to Information Retrieval, Christopher D. Manning, Prabhakar Raghavan, and Hinrich Schutze, Cambridge University Press. 2008. The links next to each lecture provide access to an online version of the text.

Some lectures have assigned readings from other papers, as shown in the link next to the lecture.

Online access to some materials is restricted to the .cmu.edu domain. CMU people can get access from outside .cmu.edu (e.g., from home) using CMU's WebVPN Service.
Recorded Lectures: Recorded lectures are available via the Heinz College video catalog. An Andrew id is required.
Homework: 3 assignments that give hands-on experience with techniques discussed in class.
Grading: 3 assignments (3 x 25%) and a final exam (25%).
Grading Scale: Grades are assigned using a curve.
Course policies: Attendance, Auditing, Laptops & mobile devices, Late homework, Pass/fail, Plagiarism & cheating, Waitlist
Syllabus (subject to revision):
Date Topic Reading
Mar 4 Course overview and introduction to text analytics (pdf)
Exploratory analysis: Frequency and co-occurrence (pdf)
HW1 out
 
Mar 11 Text representation: Turning text into features (pdf)
Exploratory analysis: Clustering (pdf)
Ch 2.0 - 2.2
Ch 16 and Ch 17  
Mar 18 Exploratory analysis: Topic models (pdf)
Predictive analysis: Categorization (pdf)
HW1 due, HW2 out
Wikipedia: Topic model, LDA
Ch 14.0-14.3
Mar 25 Predictive analysis: Categorization (pdf)
Predictive analysis: Categorization (pdf)
Ch 13
Ch 15.0-15.2
Apr 1 Predictive analysis: Categorization (pdf)
Predictive analysis: Sentiment analysis (pdf)
HW2 due, HW3 out
Ch 15.3
Feldman
Apr 8 Predictive analysis: Sentiment analysis (pdf)
Case studies: Expert finding (pdf), E-Discovery (pdf)
HW3 due
 
Apr 11 Case studies: E-Discovery (pdf); Final Review (pdf)  
Apr 15 Final exam (2:00-3:40pm) Sample final 1,
Sample final 2,
HW3 answer #8
 
 

Copyright 2016, Carnegie Mellon University.
Updated on March 08, 2016
Jamie Callan