807 - TEXT ANALYTICS Massimo Poesio Lecture 7: Wikipedia for Text Analytics.
Text Analytics Industry Use Cases
Transcript of Text Analytics Industry Use Cases
![Page 1: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/1.jpg)
Bangalore, India
[email protected] Title
Text Analytics Industry Use Cases (& the Path Forward for Text Analytics)
Aiaioo Labs - 2012
![Page 2: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/2.jpg)
• Cohan
–10 years in industry
–Research interests: NLP and ML
• Sumukh
–8 years in industry
–PhD from University of Melbourne
• Madhulika
–6 years in industry
–MS from UT Austin
–Internships at Microsoft and RRI
The Team
![Page 3: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/3.jpg)
1: AI Algorithms
2: Text Analytics Technology
3: Business Use Cases
We Develop
[email protected] What we do
![Page 4: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/4.jpg)
Bangalore, India
[email protected] Text Analytics
Research on AI Algorithms
Artificial Intelligence
Tools for Education
Tools for Graph Analysis
Tools for Text Analytics
![Page 7: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/7.jpg)
What is Text Analytics ?
Insights from Text
Meaning from Text
![Page 8: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/8.jpg)
[email protected] Use Cases
Business Use Cases
All that tech talk is fine, but how can you make us a heap more money next month?
![Page 9: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/9.jpg)
Two Ways of Using Text Analytics
Operations
Decision Making
![Page 10: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/10.jpg)
Operations - Client Use Case
![Page 11: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/11.jpg)
Decision Making - Use Case
![Page 12: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/12.jpg)
[email protected] Text Analytics
Text Analytics Details
How to handle all kinds of things people say
![Page 13: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/13.jpg)
Text Analytics
IR
NLP
Linguistics
Statistics
Psychology Study of Language Computation
Hypothesis testing
Countable
Bayesian inference
A | BCD
Social Media Documents WWW
Grammar based
ML
Discriminative
Bayesian
If - then
What is Text Analytics ?
![Page 14: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/14.jpg)
Entity Extraction
Event Extraction
Entity Extraction
Comparative Analysis
Mapping Business Needs to Text Analytics
![Page 15: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/15.jpg)
Extracting Meaning
Text: I am looking for a Hyundai car in B’lore
Syntactic Analysis: I/Pronoun am/BE looking/V for a Hyundai/NP car/NN in Bangalore/NP
Semantic Analysis: I/Person am looking for a Hyundai/Thing car/Thing in Bangalore/Place
Pragmatic Analysis: Sales Conversation
![Page 16: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/16.jpg)
Extracting Meaning
Text: I am looking for a Hyundai car in Bangalore
Semantic Analysis: I/Person am looking for a Hyundai/Thing car/Thing in Bangalore/Place
Entity Extraction
Entities: I Person Hyundai car Thing Bangalore Place
![Page 17: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/17.jpg)
Extracting Meaning
Text: Tim Cook is the new CEO of Apple Computers
Analysis: Tim/Person Cook/Person is the new CEO of Apple/Org Computers/Org
Relation Extraction
Relation: CEO_of Tim Cook (Person) Apple Computers(Org)
![Page 18: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/18.jpg)
Extracting Meaning
Raw Text: I am sad that Steve Jobs died
Analysis: This person holds a positive opinion on Steve Jobs
Sentiment Analysis
Sentiment Holder: I Object of Sentiment: Steve Jobs Polarity of Sentiment: positive
![Page 19: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/19.jpg)
Beyond Traditional Approaches to Extracting Meaning
Convey intention: I want to buy a computer.
Information: There was heavy snowfall in Sikkim.
Two Kinds of Sentences
![Page 20: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/20.jpg)
Beyond Traditional Approaches to Extracting Meaning
Raw Text: Are you sad that Steve Jobs died?
Analysis: This person is inquiring about someone’s emotions concerning Steve Jobs
Intention Analysis
Intention Holder: I Intention: inquire
![Page 21: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/21.jpg)
[email protected] Categories of Intentions
Intention Analysis – Categories of Intention
Categories Parent Category Department Urgency Source
Purchase Sales High CRM
Sell Procurement Medium ERP
Inquire Help/Sales High CRM
Direct Operations High CRM
Compare Market Research Low CRM
Suggest Market Research Low Social
Opine Design Low Social
Praise Opine Design Low Social
Criticize Opine Design Low Social
Complain Customer Service High CRM/Social
Accuse Customer Service Critical CRM
Quit Customer Service Critical CRM/Social
Express Call Center Training Low Transcripts
Thank Express Call Center Training Low Transcripts
Apologize Express Call Center Training Medium Transcripts
Empathize Express Call Center Training Medium Transcripts
![Page 22: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/22.jpg)
Approaches to Extracting Meaning
Raw Text: There is heavy snowfall in Sikkim.
Analysis: Snowfall event
Event Analysis
Event: snowfall
![Page 23: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/23.jpg)
[email protected] Categories of Events
Event Analysis
Categories
Acquisition
Merger
Spin Off
Sale
Partnership Formation
Declaration of Bankruptcy
Renaming
Closing Down
Opening Facility
Closing Facility
Business Deal
Product Launch
Product Withdrawal
Employee Joining
Employee Resignation
Employee Change of Position
![Page 24: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/24.jpg)
Approaches to Extracting Meaning
Raw Text: Bangalore is the capital of K’taka
Analysis: capital_of relation exists
Fact Analysis
Entity: Bangalore/Place Karnataka/Place Relation: Bangalore capital_of K’taka
![Page 25: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/25.jpg)
[email protected] Sentiment Analysis
Uses
Pulling report from CRM tools on loyalty, competition, etc. Computation of metrics.
Decision Making - Reports
Operations
Intention Analysis in customer service, online reputation management & placement of ads or in Alerting Systems.
![Page 26: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/26.jpg)
[email protected] Sentiment Analysis
Decision Making Uses
For strategy decisions – politics, launches
Tracking Large Numbers of Users
Metrics
Dials for navigating by
![Page 27: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/27.jpg)
[email protected] Competitive Analysis
Decision Making Example – Using Quit Intention
Quit message chart for Facebook during the week after Google+ launched.
![Page 28: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/28.jpg)
Types of Reports
1. Time Series Graphs 2. Charts 3. Tag Clouds
• Identifies popular topics 4. Event Summaries 5. Sentiment Report 6. Dials – Complaint Metrics
![Page 29: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/29.jpg)
[email protected] Sentiment Analysis
I deny that [ it can never [ be said that this is not [ a
beautiful car ] ] ].
Negative
Metrics Example – CSAT
![Page 30: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/30.jpg)
[email protected] Sentiment Analysis
Jane believes that John and not Bruce is very handsome.
Positive
Example – Sentiment Analysis – Entity Level Sentiment
Negative
About what/whom is the opinion?
![Page 31: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/31.jpg)
Section Problem Compensation
Accuracy •Low accuracy
•Trade off recall for precision
Precision •Low Precision
•Use ratios and timelines
Domain •Different domain
•Retraining
Error Compensation!
![Page 32: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/32.jpg)
tn
fn tp fp
actual target
What was selected
Precision (tp/tp+fp)
![Page 33: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/33.jpg)
tn
fn tp fp
actual target
What was selected
Recall (tp/tp+fn)
![Page 34: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/34.jpg)
Example
Precision = {inquire=0.81, direct=0.5163, accuse=0.6944, wish=0.918, compare=0.851, sell=0.9621, complain=0.8622} Recall = {direct=0.6, inquire=0.3078, accuse=0.5102, wish=0.6021, compare=0.4145, sell=0.4069, complain=0.5853}
![Page 35: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/35.jpg)
Example
F-Score = {direct=0.5566, inquire=0.4993, accuse=0.5952, wish=0.7435, compare=0.5939, sell=0.6257, complain=0.7104}
![Page 36: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/36.jpg)
What you can learn about a brand:
1: Who are a product’s strongest competitors?
2: How commoditized is the market?
3: What are the weaknesses of the product?
4: How loyal are customers in an industry?
[email protected] Strategic Market Analysis
![Page 37: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/37.jpg)
Measurement Proxy:
Use opine intention
CSAT =
number of positive mentions of brand / number of opinionated mentions of brand
Example – 1: How do you measure CSAT?
[email protected] Strategic Market Analysis
![Page 38: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/38.jpg)
Measurement Proxy:
Use compare intention
competitor’s strength =
number of mentions of competitor / number of mentions of all brands or generic products
Example – 2: Who are a product’s competitors?
[email protected] Strategic Market Analysis
![Page 39: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/39.jpg)
Measurement Proxy:
Use purchase intention
commoditization =
number of mentions of brands / number of mentions of generic product
Example – 3: How commoditized is the market?
[email protected] Strategic Market Analysis
![Page 40: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/40.jpg)
Measurement Proxy:
Use quit intention
inverse of loyalty =
number of quit intentions / total mentions
Example – 4: How loyal are customers?
[email protected] Strategic Market Analysis
![Page 41: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/41.jpg)
Measurement Proxy:
Use purchase intention
desirability =
number of purchase intentions / total mentions
Example – 5: What is the desirability of a brand?
[email protected] Strategic Market Analysis
![Page 42: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/42.jpg)
Measurement Proxy:
Use complain intention
reliability =
number of complain intentions / total mentions
Example – 6: How reliable is a product?
[email protected] Strategic Market Analysis
![Page 43: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/43.jpg)
[email protected] Sentiment Analysis
Operations Uses
For investing
Timeliness
Prioritization / Routing
Dealing with large numbers
![Page 44: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/44.jpg)
Call Center Customer Churn Model
The customer lodges a complaint
An accusation may result
Signal of intention to leave.
[email protected] Introduction – Escalation
![Page 45: Text Analytics Industry Use Cases](https://reader031.fdocuments.in/reader031/viewer/2022020702/61fb1a642e268c58cd5a2f40/html5/thumbnails/45.jpg)
Call Center Customer Sales Opportunity Model
The customer complains about a pain point (optional)
Inquires about a product feature
Signals intention to purchase or upgrade
[email protected] Introduction – Escalation