Data mining

1,097 views

Published on

0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,097
On SlideShare
0
From Embeds
0
Number of Embeds
3
Actions
Shares
0
Downloads
22
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Data mining

  1. 1. Ahmed Moussa<br />30-Feb-2010<br />
  2. 2. What is Data Mining?<br />
  3. 3. The Evolution of Data Analysis<br />
  4. 4. - <br />- RDBMS: A relational database management system<br />- ODBC: Open Database Connectivity (ODBC) provides a standard software API method for using database management systems (DBMS). <br />- OLAP: Online analytical processing, is an approach to quickly answer multi-dimensional analytical queries.<br />- SPSS: Statistical Package for the Social Sciences (formerly SPSS) is a computer program used for statistical analysis. Before 2009 it was called SPSS, but in 2009 it was re-branded as PASW.<br />
  5. 5. Results of Data Mining Include<br />
  6. 6. Results of Data Mining Include<br /><ul><li>Forecasting what may happen in the future
  7. 7. Classifying people or things into groups by recognizing patterns.
  8. 8. Clustering people or things into groups based on their attributes.
  9. 9. Associating what events are likely to occur together.
  10. 10. Sequencing what events are likely to lead to later events.</li></li></ul><li>Results of Data Mining Include<br /><ul><li>Forecasting what may happen in the future
  11. 11. Classifying people or things into groups by recognizing patterns.
  12. 12. Clustering people or things into groups based on their attributes.
  13. 13. Associating what events are likely to occur together.
  14. 14. Sequencing what events are likely to lead to later events.</li></li></ul><li>Results of Data Mining Include<br /><ul><li>Forecasting what may happen in the future
  15. 15. Classifying people or things into groups by recognizing patterns.
  16. 16. Clustering people or things into groups based on their attributes.
  17. 17. Associating what events are likely to occur together.
  18. 18. Sequencing what events are likely to lead to later events.</li></li></ul><li>Results of Data Mining Include<br /><ul><li>Forecasting what may happen in the future
  19. 19. Classifying people or things into groups by recognizing patterns.
  20. 20. Clustering people or things into groups based on their attributes.
  21. 21. Associating what events are likely to occur together.
  22. 22. Sequencing what events are likely to lead to later events.</li></li></ul><li>Results of Data Mining Include<br /><ul><li>Forecasting what may happen in the future
  23. 23. Classifying people or things into groups by recognizing patterns.
  24. 24. Clustering people or things into groups based on their attributes.
  25. 25. Associating what events are likely to occur together.
  26. 26. Sequencing what events are likely to lead to later events.</li></li></ul><li>Data mining is not<br /><ul><li>Crunching of bulk data
  27. 27. “Blind” application of algorithms
  28. 28. Going to find relationships where none exist
  29. 29. Presenting data in different ways
  30. 30. A database intensive task
  31. 31. A difficult to understand technology requiring an advanced degree in computer science</li></li></ul><li>Data Mining Is<br /><ul><li>A class of techniques that find patterns in data.
  32. 32. A user-centric, interactive process which leverages analysis technologies and computing power.
  33. 33. A group of techniques that find relationships that have not previously been discovered.
  34. 34. Not reliant on an existing database.
  35. 35. A relatively easy task that requires knowledge of the business problem/subject matter expertise.</li></li></ul><li>Data Mining Is<br /><ul><li>A class of techniques that find patterns in data.
  36. 36. A user-centric, interactive process which leverages analysis technologies and computing power.
  37. 37. A group of techniques that find relationships that have not previously been discovered.
  38. 38. Not reliant on an existing database.
  39. 39. A relatively easy task that requires knowledge of the business problem/subject matter expertise.</li></li></ul><li>Data Mining Is<br /><ul><li>A class of techniques that find patterns in data.
  40. 40. A user-centric, interactive process which leverages analysis technologies and computing power.
  41. 41. A group of techniques that find relationships that have not previously been discovered.
  42. 42. Not reliant on an existing database.
  43. 43. A relatively easy task that requires knowledge of the business problem/subject matter expertise.</li></li></ul><li>Data Mining Is<br /><ul><li>A class of techniques that find patterns in data.
  44. 44. A user-centric, interactive process which leverages analysis technologies and computing power.
  45. 45. A group of techniques that find relationships that have not previously been discovered.
  46. 46. Not reliant on an existing database.
  47. 47. A relatively easy task that requires knowledge of the business problem/subject matter expertise.</li></li></ul><li>Data Mining Is<br /><ul><li>A class of techniques that find patterns in data.
  48. 48. A user-centric, interactive process which leverages analysis technologies and computing power.
  49. 49. A group of techniques that find relationships that have not previously been discovered.
  50. 50. Not reliant on an existing database.
  51. 51. A relatively easy task that requires knowledge of the business problem/subject matter expertise.</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  52. 52. Isolate the factors that lead to fraud, waste and abuse
  53. 53. Target auditing and investigative efforts more effectively
  54. 54. Credit/Risk Scoring
  55. 55. Intrusion detection
  56. 56. Parts failure prediction
  57. 57. Recruiting/Attracting customers
  58. 58. Maximizing profitability (cross selling, identifying profitable customers)
  59. 59. Service Delivery and Customer Retention
  60. 60. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  61. 61. Isolate the factors that lead to fraud, waste and abuse
  62. 62. Target auditing and investigative efforts more effectively
  63. 63. Credit/Risk Scoring
  64. 64. Intrusion detection
  65. 65. Parts failure prediction
  66. 66. Recruiting/Attracting customers
  67. 67. Maximizing profitability (cross selling, identifying profitable customers)
  68. 68. Service Delivery and Customer Retention
  69. 69. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  70. 70. Isolate the factors that lead to fraud, waste and abuse
  71. 71. Target auditing and investigative efforts more effectively
  72. 72. Credit/Risk Scoring
  73. 73. Intrusion detection
  74. 74. Parts failure prediction
  75. 75. Recruiting/Attracting customers
  76. 76. Maximizing profitability (cross selling, identifying profitable customers)
  77. 77. Service Delivery and Customer Retention
  78. 78. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  79. 79. Isolate the factors that lead to fraud, waste and abuse
  80. 80. Target auditing and investigative efforts more effectively
  81. 81. Credit/Risk Scoring
  82. 82. Intrusion detection
  83. 83. Parts failure prediction
  84. 84. Recruiting/Attracting customers
  85. 85. Maximizing profitability (cross selling, identifying profitable customers)
  86. 86. Service Delivery and Customer Retention
  87. 87. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  88. 88. Isolate the factors that lead to fraud, waste and abuse
  89. 89. Target auditing and investigative efforts more effectively
  90. 90. Credit/Risk Scoring
  91. 91. Intrusion detection
  92. 92. Parts failure prediction
  93. 93. Recruiting/Attracting customers
  94. 94. Maximizing profitability (cross selling, identifying profitable customers)
  95. 95. Service Delivery and Customer Retention
  96. 96. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  97. 97. Isolate the factors that lead to fraud, waste and abuse
  98. 98. Target auditing and investigative efforts more effectively
  99. 99. Credit/Risk Scoring
  100. 100. Intrusion detection
  101. 101. Parts failure prediction
  102. 102. Recruiting/Attracting customers
  103. 103. Maximizing profitability (cross selling, identifying profitable customers)
  104. 104. Service Delivery and Customer Retention
  105. 105. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  106. 106. Isolate the factors that lead to fraud, waste and abuse
  107. 107. Target auditing and investigative efforts more effectively
  108. 108. Credit/Risk Scoring
  109. 109. Intrusion detection
  110. 110. Parts failure prediction
  111. 111. Recruiting/Attracting customers
  112. 112. Maximizing profitability (cross selling, identifying profitable customers)
  113. 113. Service Delivery and Customer Retention
  114. 114. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br /><ul><li>Fraud/Non-Compliance Anomaly detection
  115. 115. Isolate the factors that lead to fraud, waste and abuse
  116. 116. Target auditing and investigative efforts more effectively
  117. 117. Credit/Risk Scoring
  118. 118. Intrusion detection
  119. 119. Parts failure prediction
  120. 120. Recruiting/Attracting customers
  121. 121. Maximizing profitability (cross selling, identifying profitable customers)
  122. 122. Service Delivery and Customer Retention
  123. 123. Build profiles of customers likely to use which services</li></li></ul><li>Examples of What People are Doing with Data Mining:<br />Right offer for the right customer throw the right channel in the right time<br />
  124. 124. How Can We Do Data Mining?<br /><ul><li>A standard process
  125. 125. Existing data
  126. 126. Software technologies
  127. 127. Situational expertise</li></li></ul><li>How Can We Do Data Mining?<br /><ul><li>A standard process
  128. 128. Existing data
  129. 129. Software technologies
  130. 130. Situational expertise</li></ul>The data mining process must be reliable and repeatable by people with little data mining background.<br />
  131. 131. Phases and Tasks<br />Business<br />Understanding<br />Data<br />Understanding<br />Data<br />Preparation<br />Modeling<br />Deployment<br />Evaluation<br />Determine <br /> Business Objectives<br />Background<br />Business Objectives<br />Business Success <br /> Criteria<br />Situation Assessment<br />Inventory of Resources<br />Requirements,<br /> Assumptions, and<br /> Constraints<br />Risks and Contingencies<br />Terminology<br />Costs and Benefits<br />Determine <br /> Data Mining Goal<br />Data Mining Goals<br />Data Mining Success <br /> Criteria<br />Produce Project Plan<br />Project PlanInitial Asessment of Tools and Techniques<br />Collect Initial Data<br />Initial Data Collection <br /> Report<br />Describe Data<br />Data Description Report<br />Explore Data<br />Data Exploration Report Verify Data Quality <br />Data Quality Report<br />Data Set<br />Data Set Description<br />Select Data <br />Rationale for Inclusion / <br /> Exclusion<br />Clean Data <br />Data Cleaning Report<br />Construct Data<br />Derived Attributes<br />Generated Records<br />Integrate Data<br />Merged Data<br />Format Data<br />Reformatted Data<br />Select Modeling<br /> Technique<br />Modeling Technique<br />Modeling Assumptions<br />Generate Test Design<br />Test Design<br />Build Model<br />Parameter Settings<br />Models<br />Model Description<br />Assess Model<br />Model AssessmentRevised Parameter Settings<br />Evaluate Results<br />Assessment of Data <br /> Mining Results w.r.t. <br /> Business Success <br /> Criteria<br />Approved Models<br />Review Process<br />Review of Process<br />Determine Next Steps<br />List of Possible Actions<br />Decision<br />Plan Deployment<br />Deployment Plan<br />Plan Monitoring and <br /> Maintenance<br />Monitoring and Maintenance Plan<br />Produce Final Report<br />Final Report<br />Final Presentation<br />Review Project<br />Experience <br /> Documentation<br />
  132. 132. Phases and Tasks<br />
  133. 133. Phases and Tasks<br />
  134. 134. Phases and Tasks<br />A) Business Understanding<br />Determine Business Objectives<br />Background<br />Business Objectives<br />Business Success <br />Criteria<br />Situation Assessment<br />Inventory of Resources <br />Requirements, Assumptions, and Constraints<br />Risks and Contingencies<br />Terminology<br />Costs and Benefits<br />Determine Data Mining Goal <br />Data Mining Goals<br />Data Mining Success <br /> Criteria<br />Produce Project Plan<br />Project PlanInitial Asessment of Tools and Techniques<br />
  135. 135. Phases and Tasks<br />
  136. 136. Phases and Tasks<br />B) Data Understanding<br />Explore Data<br />Data Exploration Report Verify Data Quality <br />Data Quality Report<br />Collect Initial Data<br />Initial Data Collection <br />Report<br />Describe Data<br />Data Description Report<br />
  137. 137. Phases and Tasks<br />
  138. 138. Phases and Tasks<br />C) Data Preparation<br />Data Set<br />Data Set Description<br />Select Data <br />Rationale for Inclusion/Exclusion<br />Clean Data <br />Data Cleaning Report<br />Integrate Data<br />Merged Data<br />Format Data<br />Reformatted Data<br />Construct Data<br />Derived Attributes<br />Generated Records<br />
  139. 139. Phases and Tasks<br />
  140. 140. Phases and Tasks<br />D) Modeling<br />Select Modeling<br />Modeling Technique<br />Modeling Assumptions<br />Generate Test Design<br />Test Design<br />Build Model<br />Parameter Settings<br />Models and Model Description<br />Assess Model<br />Model AssessmentRevised Parameter Settings<br />
  141. 141. Phases and Tasks<br />
  142. 142. Phases and Tasks<br />D) Evaluation<br />Evaluate Results<br />Assessment of Data <br /> Mining Results w.r.t. <br /> Business Success <br /> Criteria<br />Approved Models<br />Review Process<br />Review of Process<br />Determine Next Steps<br />List of Possible Actions<br />Decision<br />
  143. 143. Phases and Tasks<br />
  144. 144. Phases and Tasks<br />E) Deployment<br />Plan Deployment<br />Deployment Plan<br />Plan Monitoring and <br />Maintenance<br />Monitoring and Maintenance Plan<br />Produce Final Report<br />Final Report<br />Final Presentation<br />Review Project<br />Experience and Documentation<br />
  145. 145. Data mining success story <br />The US Internal Revenue Service <br />needed to improve customer service and...<br />Scheduled its workforce <br />to provide faster, more accurate answers to questions.<br />
  146. 146. Data mining success story <br />The US Drug Enforcement Agency needed to be more effective in their drug “busts” and<br />analyzed suspects’ cell phone usage to focus investigations.<br />
  147. 147. Data mining success story <br />HSBC need to cross-sell more effectively by identifying profiles that would be interested in higher yielding investments and...<br />Reduced direct mail costs by 30% while garnering 95% of the campaign’s revenue.<br />
  148. 148. Final Comments<br />
  149. 149. <ul><li>Data Mining can be utilized in any organization that needs to find patterns or relationships in their data.</li></li></ul><li><ul><li>Data Mining can be utilized in any organization that needs to find patterns or relationships in their data.
  150. 150. By using the DM methodology, analysts can have a reasonable level of assurance that their Data Mining efforts will render useful, repeatable, and valid results.</li></li></ul><li>

×