Analysis types, Applications you plan to do more in 2008

Poll

Which analysis types / applications you plan to do more in 2008 (vs 2007): [170 voters] Analysis type Classification (91) 53.5% Clustering/ Segmentation (83) 48.8% Regression/ Forecasting (66) 38.8% Association rules (61) 35.9% Text mining (40) 23.5% Visualization (35) 20.6% Time sequences (35) 20.6% Web analytics (32) 18.8% Link analysis/ Social networks (30) 17.6% None (8) 4.7% Other (7) 4.1% Application area CRM/ consumer analytics (56) 32.9% Finance / investments (35) 20.6% Fraud detection (27) 15.9% Biotech/ Medical/ Pharma (25) 14.7% Advertising/ecommerce (22) 12.9% Web usage mining (21) 12.4% Science (21) 12.4% Web content mining/search (16) 9.4% Health care/ Insurance (16) 9.4% Telecom / Cable (15) 8.8% Security / Anti-terrorism (8) 4.7% Junk email/ anti-spam (5) 2.9% Other applications (16) 9.4% No app areas (33) 19.4%

Notes About 95% of respondents were planning to use some new methods, and the average number of new methods was almost three (2.96).

About 80% of respondents had new applications, and the average number of new applications was close to two (2.05).

Here are top analysis types selected among respondents who chose these application areas. Note that Clustering is most associated with CRM and Fraud detection, while classification with Finance/Investments.

• CRM/ consumer analytics: (avg 3.8 analysis types)
  1. Clustering/ Segmentation
  2. Classification
  3. Regression/ Forecasting
  4. Association rules
  5. Web analytics
  
  
• Finance / investments: (avg 3.5 analysis types)
  1. Classification
  2. Clustering/ Segmentation
  3. Regression/ Forecasting
  4. Association rules
  5. Time sequences
  
  
• Fraud detection: (avg 3.6 analysis types)
  1. Clustering/ Segmentation
  2. Classification
  3. Regression/ Forecasting
  4. Time sequences
  5. Text mining
  
  
• Biotech/ Medical/ Pharma: (avg 3.2 analysis types)
  1. Regression/ Forecasting
  2. Clustering/ Segmentation
  3. Classification
  4. Association rules
  5. Time sequences
  
  
• Advertising/ecommerce: (avg 4.3 analysis types)
  1. Clustering/ Segmentation
  2. Classification
  3. Web analytics
  4. Text mining
  5. Association rules

Here are application areas most associated with top analysis types.

• Classification - most associated applications:
  1. CRM/ consumer analytics
  2. Finance / investments
  3. Fraud detection
  4. Science
  5. Advertising/ecommerce
  
  
• Clustering/ Segmentation - most associated applications
  1. CRM/ consumer analytics
  2. Finance / investments
  3. Fraud detection
  4. Advertising/ecommerce
  5. Biotech/ Medical/ Pharma
  
  
• Regression/ Forecasting - most associated applications
  1. CRM/ consumer analytics
  2. Fraud detection
  3. Finance / investments
  4. Biotech/ Medical/ Pharma
  5. Science
  
  
• Association rules - most associated applications
  1. CRM/ consumer analytics
  2. Finance / investments
  3. Biotech/ Medical/ Pharma
  4. no application
  5. Fraud detection
  
  
• Text mining - most associated applications
  1. CRM/ consumer analytics
  2. Advertising/ecommerce
  3. Web content mining/search
  4. Web usage mining
  5. Fraud detection