Machine Learning (ML) Classification

1
Machine Learning (ML) Classification

• Tim Humphrey
• LexisNexis
• 21 June 2001

2
A little scientific humor

• This atom says to his friend, "I'm really upset,
  I've just lost an electron."
• His friend says to him, "Are you sure?"
• "Yeah," he replies. "I'm Positive.
• __________________________________________________
  __________________________
• How many weeks are there in a light year?

3
Overview

• Preface
• Definition of ML Classifier
• You as an ML Classifier
• A little math
• Examples of ML algorithms
• How well do ML classifiers work?
• Advantages Disadvantages
• Uses of ML classifiers
• Challenges
• Q A

4
Preface

• Mark talked about ways to make rules that
  classify documents. Examples of companies that
  have such systems are
• LexisNexis
• Verity
• SmartLogic
• Interwoven
• Machine Learning is another way of getting
  computers to classify documents. Machine learning
  is normally not rule based. Instead, it is
  normally statistically based.

5
Definition of ML Classifier

• Definition of Machine Learning from
  dictionary.com
• The ability of a machine to improve its
  performance based on previous results.
• So, machine learning document classification is
  the ability of a machine to improve its document
  classification performance based on previous
  results of document classification.

6
You as an ML Classifier

• Topic 1 words
• baseball, owners, sports, selig, ball, bill,
  indians, isringhausen, mets, minors, players,
  specter, stadium, power, send, new, bud, comes,
  compassion, game, headaches, lite, nfl, powerful,
  strawberry, urges, home, ambassadors, building,
  calendar, commish, costs, day, dolan, drive,
  hits, league, little, match, payments, pitch,
  play, player, red, stadiums, umpire, wife, youth,
  field, leads
• Topic 2 words
• merger, business, bank, buy, announces, new,
  acquisition, finance, companies, com, company,
  disclosure, emm, news, us, acquire, chemical,
  inc, results, shares, takeover, corporation,
  european, financial, investment, market, quarter,
  two, acquires, bancorp, bids, communications,
  first, mln, purchase, record, stake, west, sale,
  bid, bn, brief, briefs, capital, control, europe,
  inculab

7
Use the previous slides topics related words
to classify the following titles

1. CYBEX-Trotter merger creates fitness equipment
   powerhouse
2. WSU RECRUIT CHOOSES BASEBALL INSTEAD OF FOOTBALL
3. FCC chief says merger may help pre-empt Internet
   regulation
4. Vision of baseball stadium growing
5. Regency Realty Corporation Completes Acquisition
   Of Branch properties
6. Red Sox to punish All-Star scalpers
7. Canadian high-tech firm poised to make
   415-million acquisition
8. Futures-selling hits the Footsie for six
9. A'S NOTEBOOK Another Young Arm Called Up
10. All-American SportPark Reaches Agreement for
    Release of Corporate Guarantees

8
Titles Their Classifications

1. (2) CYBEX-Trotter merger creates fitness
   equipment powerhouse
2. (1) WSU RECRUIT CHOOSES BASEBALL INSTEAD OF
   FOOTBALL
3. (2) FCC chief says merger may help pre-empt
   Internet regulation
4. (1) Vision of baseball stadium growing
5. (2) Regency Realty Corporation Completes
   Acquisition Of Branch properties
6. (1) Red Sox to punish All-Star scalpers
7. (2) Canadian high-tech firm poised to make
   415-million acquisition
8. (2) Futures-selling hits the Footsie for six
9. (1) A'S NOTEBOOK Another Young Arm Called Up
10. (1) All-American SportPark Reaches Agreement for
    Release of Corporate Guarantees

9
The Salary Theorem

• Mathematic Proof of The less you know, the more
  you make.
• Knowledge is Power
• Time is Money
• Power Work / Time
• Knowledge Work/Money
• Solving for Money, we get
• Money Work / Knowledge.
• Thus, as Knowledge approaches zero, Money
  approaches infinity,
• regardless of the amount of work done.
• Conclusion The less you know, the more you make.

10
A little math

• Canadian high-tech firm poised to make
  415-million acquisition
• Estimate the probablity of a word in a topic by
  dividing the number of times the word appeared in
  the topics training set by the total number of
  word occurrences in the topics training set.
• For each topic,T, sum the probability of finding
  each word of the title in a title that is
  classified as T.
• The title is classified as the topic with the
  largest sum.
• Titles evidence of being in Topic 20.01152
• Titles evidence of being in Topic 10.00932
• Canadian 1 0 high 0 0 tech 2 0 firm 1 0
• poised 0 0 make 0 0 million 4 4
• acquisition 10 0
• of words in Topic2 1563
• of words in Topic1 429

11
Examples of ML algorithms

• Naïve Bayes This method computes the
  probability that a document is about a particular
  topic, T, using a) the words of the document to
  be classified and b) the estimated probability of
  each of these words as they appeared in the set
  of training documents for the topic, T like the
  example previously given.
• Neural networks During training, a neural
  network looks at the patterns of features (e.g.
  words, phrases, or N-grams) that appear in a
  document of the training set and attempts to
  produce classifications for the document. If its
  attempt doesnt match the set of desired
  classifications, it adjusts the weights of the
  connections between neurons. It repeats this
  process until the attempted classifications match
  the desired classifications.
• Instance based Saves documents of the training
  set and compares new documents to be classified
  with the saved documents. The document to be
  classified gets tagged with the highest scoring
  classifications. One way to do this is to
  implement a search engine using the documents of
  the training set as the document collection. A
  document to be classified becomes a query/search.
  A classification, C, is picked if a large number
  of its training set documents are at the top of
  the returned answer set.

12
How well do ML classifiers work?

• A good system will have an accuracy of above 80.
• Strong evidence of how good these systems are is
  the number of companies in the market place with
  machine learning document classification systems.
  
• Example are Semio, Inxight, Purple Yogi,
  Hummingbird, Autonomy, 80-20 Inc., Dophin Search,
  Textology Inc.,

13
Advantages Disadvantages

• Advantage over classification by humans Once the
  system is trained, classification is done
  automatically with no or little human
  intervention saving human resources.
• Advantage over classification by humans
  Consistent classification.
• Advantage over rule based classification Human
  resources are not needed to make rules.
• Disadvantage Not always obvious why it
  classified a document in a certain way and not
  obvious how to keep it from doing the same type
  of classification in the future (i.e. dont know
  how to modify it.)
• Disadvantage Human resources must be used to
  manually classify documents for the training set.
  Furthermore, the number and type of document that
  should be in the training set isnt
  straightforward.

14
Uses of ML classifiers

• Automatically classify documents.
• Suggest classifications that a human can pick
  from.
• Classify paragraphs or even sentences.
• Find important information in a document. For
  example, rules of law in a case law document, or
  the facts of the case.

15
Challenges

• Labeling the documents of the training set.
• What is the best way to pick documents for the
  training set so the machine-learning algorithm
  produces a classifier with high accuracy?
• Which machine-learning algorithm works best on
  your classification problem?

16
Questions Answers