There are many examples of text-based documents (all in ‘electronic’ format…) e-mails, corporate Web pages, customer surveys, résumés, medical records, DNA sequences, technical papers, incident reports, news stories and more… Not enough time or patience to read Can we extract the most vital kernels of information… So, we wish to find a way to gain knowledge (in summarised form) from all that text, without reading or examining them fully first…! Some others (e.g. DNA seq.) are hard to comprehend!

1. Text Mining
By
SATHISHKUMAR G
(sathishsak111@gmail.com)

2. What is Text Mining?
• There are many examples of text-based documents (all in
‘electronic’ format…)
– e-mails, corporate Web pages, customer surveys, résumés, medical records,
DNA sequences, technical papers, incident reports, news stories and
more…
• Not enough time or patience to read
– Can we extract the most vital kernels of information…
• So, we wish to find a way to gain knowledge (in summarised form)
from all that text, without reading or examining them fully first…!
– Some others (e.g. DNA seq.) are hard to comprehend!

3. What is Text Mining?
• Traditional data mining uses ‘structured data’ (n x p
matrix)
• The analysis of ‘free-form text’ is also referred to as
‘unstructured data’,
– successful categorisation of such data can be a difficult and
time-consuming task…
• Often, can combine free-form text and structured data to
derive valuable, actionable information… (e.g. as in
typical surveys) – semi-structured

4. Text Mining: Examples
• Text mining is an exercise to gain knowledge from stores
of language text.
• Text:
– Web pages
– Medical records
– Customer surveys
– Email filtering (spam)
– DNA sequences
– Incident reports
– Drug interaction reports
– News stories (e.g. predict stock movement)

5. What is Text Mining
• Data examples
– Web pages
– Customer surveys
Customer Age Sex Tenure Comments Outcome
123 24 M 12 years Incorrect
charges on bill
customer angry
Y
243 26 F 1 month Inquiry about
charges to India
N
346 54 M 3 years Question about
charges on bill
N

6. Amazon.com

7. Of Mice and Men: Concordance
Concordance is an alphabetized list of the most frequently occurring words in a book,
excluding common words such as "of" and "it." The font size of a word is proportional to the
number of times it occurs in the book.

8. Of Mice and Men: Text Stats

9. Text Mining: Yahoo Buzz

10. Text Mining: Google News

11. Text Mining
• Typically falls into one of two categories
– Analysis of text: I have a bunch of text I am
interested in, tell me something about it
• E.g. sentiment analysis, “buzz” searches
– Retrieval: There is a large corpus of text documents,
and I want the one closest to a specified query
• E.g. web search, library catalogs, legal and medical
precedent studies

12. Text Mining: Analysis
• Which words are most present
• Which words are most surprising
• Which words help define the document
• What are the interesting text phrases?

13. Text Mining: Retrieval
• Find k objects in the corpus of documents which
are most similar to my query.
• Can be viewed as “interactive” data mining -
query not specified a priori.
• Main problems of text retrieval:
– What does “similar” mean?
– How do I know if I have the right documents?
– How can I incorporate user feedback?

14. Text Retrieval: Challenges
• Calculating similarity is not obvious - what is the distance between
two sentences or queries?
• Evaluating retrieval is hard: what is the “right” answer ? (no
ground truth)
• User can query things you have not seen before e.g. misspelled,
foreign, new terms.
• Goal (score function) is different than in classification/regression:
not looking to model all of the data, just get best results for a
given user.
• Words can hide semantic content
– Synonymy: A keyword T does not appear anywhere in the document, even
though the document is closely related to T, e.g., data mining
– Polysemy: The same keyword may mean different things in different
contexts, e.g., mining

15. Basic Measures for Text Retrieval
• Precision: the percentage of retrieved documents that are in
fact relevant to the query (i.e., “correct” responses)
• Recall: the percentage of documents that are relevant to the
query and were, in fact, retrieved
recall=
|{Relevant}∩{Retrieved}|
|{Relevant}|
|}{|
|}{}{|
Retrieved
RetrievedRelevant
precision
∩
=

16. Precision vs. Recall
• We’ve been here before!
– Precision = TP/(TP+FP)
– Recall = TP/(TP+FN)
– Trade off:
• If algorithm is ‘picky’: precision high, recall low
• If algorithm is ‘relaxed’: precision low, recall high
– BUT: recall often hard if not impossible to calculate
Truth:Relvant Truth:Not Relevant
Algorithm:Relevant TP FP
Algorithm: Not Relevant FN TN
16
1 0
1 a b
0 c d
predicted
outcome
actual
outcome

17. Precision Recall Curves
• If we have a labelled training set, we can calculate recall.
• For any given number of returned documents, we can plot a point
for precision vs. recall. (similar to thresholds in ROC curves)
• Different retrieval algorithms might have very different curves -
hard to tell which is “best”

18. Term / document matrix
• Most common form of representation in text
mining is the term - document matrix
– Term: typically a single word, but could be a word
phrase like “data mining”
– Document: a generic term meaning a collection of
text to be retrieved
– Can be large - terms are often 50k or larger,
documents can be in the billions (www).
– Can be binary, or use counts

19. Term document matrix
• Each document now is just a vector of terms, sometimes
boolean
Database SQL Index Regression Likelihood linear
D1 24 21 9 0 0 3
D2 32 10 5 0 3 0
D3 12 16 5 0 0 0
D4 6 7 2 0 0 0
D5 43 31 20 0 3 0
D6 2 0 0 18 7 6
D7 0 0 1 32 12 0
D8 3 0 0 22 4 4
D9 1 0 0 34 27 25
D10 6 0 0 17 4 23
Example: 10 documents: 6 terms

20. Term document matrix
• We have lost all semantic content
• Be careful constructing your term list!
– Not all words are created equal!
– Words that are the same should be treated the same!
• Stop Words
• Stemming

21. Stop words
• Many of the most frequently used words in English are worthless in
retrieval and text mining – these words are called stop words.
– the, of, and, to, ….
– Typically about 400 to 500 such words
– For an application, an additional domain specific stop words list may be
constructed
• Why do we need to remove stop words?
– Reduce indexing (or data) file size
• stopwords accounts 20-30% of total word counts.
– Improve efficiency
• stop words are not useful for searching or text mining
• stop words always have a large number of hits

22. Stemming
• Techniques used to find out the root/stem of a word:
– E.g.,
– user engineering
– users engineered
– used engineer
– using
• stem: use engineer
Usefulness
• improving effectiveness of retrieval and text mining
– matching similar words
• reducing indexing size
– combing words with same roots may reduce indexing size as
much as 40-50%.

23. Basic stemming methods
• remove ending
– if a word ends with a consonant other than s,
followed by an s, then delete s.
– if a word ends in es, drop the s.
– if a word ends in ing, delete the ing unless the remaining word consists only of
one letter or of th.
– If a word ends with ed, preceded by a consonant, delete the ed unless this
leaves only a single letter.
– …...
• transform words
– if a word ends with “ies” but not “eies” or “aies” then “ies --> y.”

24. Feature Selection
• Performance of text classification algorithms can be optimized by
selecting only a subset of the discriminative terms
– Even after stemming and stopword removal.
• Greedy search
– Start from full set and delete one at a time
– Find the least important variable
• Can use Gini index for this if a classification problem
• Often performance does not degrade even with orders of
magnitude reductions
– Chakrabarti, Chapter 5: Patent data: 9600 patents in communcation,
electricity and electronics.
– Only 140 out of 20,000 terms needed for classification!

25. Distances in TD matrices
• Given a term doc matrix represetnation, now we can define
distances between documents (or terms!)
• Elements of matrix can be 0,1 or term frequencies (sometimes
normalized)
• Can use Euclidean or cosine distance
• Cosine distance is the angle between the two vectors
• Not intuitive, but has been proven to work well
• If docs are the same, dc =1, if nothing in common dc=0

26. • We can calculate cosine and Euclidean distance
for this matrix
• What would you want the distances to look like?
Database SQL Index Regression Likelihood linear
D1 24 21 9 0 0 3
D2 32 10 5 0 3 0
D3 12 16 5 0 0 0
D4 6 7 2 0 0 0
D5 43 31 20 0 3 0
D6 2 0 0 18 7 6
D7 0 0 1 32 12 0
D8 3 0 0 22 4 4
D9 1 0 0 34 27 25
D10 6 0 0 17 4 23

27. Document distance
• Pairwise distances between documents
• Image plots of cosine distance, Euclidean, and
scaled Euclidean
R function: ‘image’

28. Weighting in TD space
• Not all phrases are of equal importance
– E.g. David less important than Beckham
– If a term occurs frequently in many documents it has less discriminatory
power
– One way to correct for this is inverse-document frequency (IDF).
– Term importance = Term Frequency (TF) x IDF
– Nj= # of docs containing the term
– N = total # of docs
– A term is “important” if it has a high TF and/or a high IDF.
– TF x IDF is a common measure of term importance

29. Database SQL Index Regression Likelihood linear
D1 24 21 9 0 0 3
D2 32 10 5 0 3 0
D3 12 16 5 0 0 0
D4 6 7 2 0 0 0
D5 43 31 20 0 3 0
D6 2 0 0 18 7 6
D7 0 0 1 32 12 0
D8 3 0 0 22 4 4
D9 1 0 0 34 27 25
D10 6 0 0 17 4 23
Database SQL Index Regression Likelihood linear
D1 2.53 14.6 4.6 0 0 2.1
D2 3.3 6.7 2.6 0 1.0 0
D3 1.3 11.1 2.6 0 0 0
D4 0.7 4.9 1.0 0 0 0
D5 4.5 21.5 10.2 0 1.0 0
D6 0.2 0 0 12.5 2.5 11.1
D7 0 0 0.5 22.2 4.3 0
D8 0.3 0 0 15.2 1.4 1.4
D9 0.1 0 0 23.56 9.6 17.3
D10 0.6 0 0 11.8 1.4 16.0
TF IDF

30. Queries
• A query is a representation of the user’s information needs
– Normally a list of words.
• Once we have a TD matrix, queries can be represented as a vector in
the same space
– “Database Index” = (1,0,1,0,0,0)
• Query can be a simple question in natural language
• Calculate cosine distance between query and the TF x IDF version of
the TD matrix
• Returns a ranked vector of documents

31. Latent Semantic Indexing
• Criticism: queries can be posed in many ways, but still
mean the same
– Data mining and knowledge discovery
– Car and automobile
– Beet and beetroot
• Semantically, these are the same, and documents with
either term are relevant.
• Using synonym lists or thesauri are solutions, but messy
and difficult.
• Latent Semantic Indexing (LSI): tries to extract hidden
semantic structure in the documents
• Search what I meant, not what I said!

32. LSI
• Approximate the T-dimensional term space using
principle components calculated from the TD matrix
• The first k PC directions provide the best set of k
orthogonal basis vectors - these explain the most
variance in the data.
– Data is reduced to an N x k matrix, without much loss of
information
• Each “direction” is a linear combination of the input
terms, and define a clustering of “topics” in the data.
• What does this mean for our toy example?

33. Database SQL Index Regression Likelihood linear
D1 24 21 9 0 0 3
D2 32 10 5 0 3 0
D3 12 16 5 0 0 0
D4 6 7 2 0 0 0
D5 43 31 20 0 3 0
D6 2 0 0 18 7 6
D7 0 0 1 32 12 0
D8 3 0 0 22 4 4
D9 1 0 0 34 27 25
D10 6 0 0 17 4 23
Database SQL Index Regression Likelihood linear
D1 2.53 14.6 4.6 0 0 2.1
D2 3.3 6.7 2.6 0 1.0 0
D3 1.3 11.1 2.6 0 0 0
D4 0.7 4.9 1.0 0 0 0
D5 4.5 21.5 10.2 0 1.0 0
D6 0.2 0 0 12.5 2.5 11.1
D7 0 0 0.5 22.2 4.3 0
D8 0.3 0 0 15.2 1.4 1.4
D9 0.1 0 0 23.56 9.6 17.3
D10 0.6 0 0 11.8 1.4 16.0

34. LSI
• Typically done using Singular Value Decomposition
(SVD) to find principal components
TD matrix
Term weighting by
document - 10 x 6
New orthogonal basis for the
data (PC directions) -
Diagonal matrix of
eigenvalues
For our example: S = (77.4,69.5,22.9,13.5,12.1,4.8)
Fraction of the variance explained (PC1&2) = = 92.5%

35. LSI
Top 2 PC make new
pseudo-terms to define
documents…
Also, can look at first two Principal components:
(0.74,0.49, 0.27,0.28,0.18,0.19) -> emphasizes first two terms
(-0.28,-0.24,-0.12,0.74,0.37,0.31) -> separates the two clusters
Note how distance from the origin shows number of terms,
And angle (from the origin) shows similarity as well

36. LSI
• Here we show the same
plot, but with two new
documents, one with the
term “SQL” 50 times,
another with the term
“Databases” 50 times.
• Even though they have no
phrases in common, they
are close in LSI space

37. Textual analysis
• Once we have the data into a nice matrix
representation (TD, TDxIDF, or LSI), we can
throw the data mining toolbox at it:
– Classification of documents
• If we have training data for classes
– Clustering of documents
• unsupervised

38. Automatic document classification
• Motivation
– Automatic classification for the tremendous number of on-line text documents (Web
pages, e-mails, etc.)
– Customer comments: Requests for info, complaints, inquiries
• A classification problem
– Training set: Human experts generate a training data set
– Classification: The computer system discovers the classification rules
– Application: The discovered rules can be applied to classify new/unknown documents
• Techniques
– Linear/logistic regression, naïve Bayes
– Trees not so good here due to massive dimension, few interactions

39. Naïve Bayes Classifier for Text
• Naïve Bayes classifier = conditional independence model
– Also called “multivariate Bernoulli”
– Assumes conditional independence assumption given the class:
p( x | ck ) = Π p( xj | ck )
– Note that we model each term xj as a discrete random variable
In other words, the probability that a bunch of words comes from a given class
equals the product of the individual probabilities of those words.

40. jNx
j
kjkxk
n
cxpcNpcp ∏=
∝
1
)|()|()|(x
Multinomial Classifier for Text
• Multinomial Classification model
– Assumes that the data are generated by a p-sided die (multinomial model)
– where Nx = number of terms (total count) in document x
– xj = number of times term j occurs in the document
– ck = class = k
– Based on training data, each class has its own multinomial probability across all
words.
.

41. Naïve Bayes vs. Multinomial
• Many extensions and adaptations of both
• Text mining classification models usually a version of one of these
• Example: Web pages
– Classify webpages from CS departments into:
• student, faculty, course,project
– Train on ~5,000 hand-labeled web pages from Cornell, Washington,
U.Texas, Wisconsin
– Crawl and classify a new site (CMU)
Student Faculty Person Project Course Departmt
Extracted 180 66 246 99 28 1
Correct 130 28 194 72 25 1
Accuracy: 72% 42% 79% 73% 89% 100%

42. NB vs. multinomial

43. Highest Probability Terms in Multinomial Distributions
Classifying web pages at a University:

44. Document Clustering
• Can also do clustering, or unsupervised learning of docs.
• Automatically group related documents based on their
content.
• Require no training sets or predetermined taxonomies.
• Major steps
– Preprocessing
• Remove stop words, stem, feature extraction, lexical analysis, …
– Hierarchical clustering
• Compute similarities applying clustering algorithms, …
– Slicing
• Fan out controls, flatten the tree to desired number of levels.
• Like all clustering examples, success is relative

45. Document Clustering
• To Cluster:
– Can use LSI
– Another model: Latent Dirichlet Allocation (LDA)
– LDA is a generative probabilistic model of a corpus. Documents are
represented as random mixtures over latent topics, where a topic is
characterized by a distribution over words.
• LDA:
– Three concepts: words, topics, and documents
– Documents are a collection of words and have a probability
distribution over topics
– Topics have a probability distribution over words
– Fully Bayesian Model

46. LDA
• Assume data was generated by a generative process:
∀ θ is a document - made up from topics from a probability distribution
• z is a topic made up from words from a probability distribution
• w is a word, the only real observables (N=number of words in all documents)
• Then, the LDA equations are specified in a fully Bayesian model:
α=per-document topic distributions

47. Which can be solved via advance
computational techniques
see Blei, et al 2003

48. LDA output
• The result can be an often-useful classification of documents into topics, and a
distribution of each topic across words:

49. Another Look at LDA
• Model: Topics made up of words used to generate documents

50. Another Look at LDA
• Reality: Documents observed, infer topics

51. Case Study: TV Listings
• Use text to make recommendations for TV shows

52. Data Issues
• 10013|In Harm's Way|In Harm's Way|A tough Naval officer faces the enemy
while fighting in the South Pacific during World War II.|A tough Naval
officer faces the enemy while fighting in the South Pacific during World
War II.|en-US| Movie,NR Rating|Movies:Drama|||165|1965|USA||||||STARS-3||
NR|John Wayne, Kirk Douglas, Patricia Neal, Tom Tryon, Paula Prentis s,
Burgess Meredith|Otto Preminger||||Otto Preminger|
Parsed Program Guide entries – 2 weeks, ~66,000 programs, 19,000 words
• Collapse on series (syndicated shows are still a problem)
• Stopwords/stemming, duplication, paid programming, length
normalization

53. Data Processing
• Combine shows from one series into a ‘canonical’ format

54. Results
• We fit LDA
– Results in a full distribution of words, topics and documents
– Topics are unveiled which are a collection of words

55. Results
• For user modelling, consider the collection of shows a single user
watches as a ‘document’ – then look to see what topics (and hence,
words) make up that document

57. Show mining via text

58. Text Mining: Helpful Data
• WordNet
Courtesy: Luca Lanzi

59. Text Mining - Other Topics
• Part of Speech Tagging
– Assign grammatical tags to words (verb, noun, etc)
– Helps in understanding documents : uses Hidden Markov Models
• Named Entity Classification
– Classification task: can we automatically detect proper nouns and tag them
– “Mr. Jones” is a person; “Madison” is a town.
– Helps with dis-ambiguation: spears

60. Text Mining - Other Topics
• Sentiment Analysis
– Automatically determine tone in text: positive, negative or neutral
– Typically uses collections of good and bad words
– “While the traditional media is slowly starting to take John McCain’s straight talking
image with increasingly large grains of salt, his base isn’t quite ready to give up on their
favorite son. Jonathan Alter’s bizarre defense of McCain after he was caught telling an
outright lie, perfectly captures that reluctance[.]”
– Often fit using Naïve Bayes
• There are sentiment word lists out there:
– See http://neuro.imm.dtu.dk/wiki/Text_sentiment_analysis

61. Text Mining - Other Topics
• Summarizing text: Word Clouds
– Takes text as input, finds the most
interesting ones, and displays them
graphically
– Blogs do this
– Wordle.net

62. Modest Mouse lyrics

63. Thank you