Data mining
- 1. Nandini Patil Assistant Professor Godutai Engg. College,kalaburagi Data Mining
- 2. Introduction data mining is the art and science of discovering the knowledge insights and patterns in the data. it is act of extracting useful pattern from an organized Collection of data. patterns must be valid, novel, potentially useful and understandable. data mining is a multidisciplinary field that borrows techniques from a variety of field. it utilizes the knowledge of the data quality & data organizing from the databases area.
- 3. Cont… It draws- Analytical techniques from statistics and computer science knowledge of decision making from field of business management. Example: customer who buy cheese & milk also buy bread
- 4. Gathering & selecting data Growth of data is coming with higher velocity, volume & variety. to learn from Data quality data needs to be effectively gathered in the and organised and then sufficiently mined. gathering and curating data takes time & efforts when data is unstructured or semistructured. knowledge of the business domain helps to select the right streams of the data for pursuing the new insights.
- 5. Data cleaning & preparation Duplicate data needs to be removed Missing values need to be filled in Data element should be comparable Continuous values may need to be binned Outlier data elements need to be removed Ensure that the data is representative of the phenomena Data may need to be selected to increase information density
- 6. outputs of data mining data mining output servers different types of the objective data mining output are decision tree regression evaluation or mathematical functions Some business rules
- 7. Evaluating data mining Results predictive accuracy = correct predictions /Total predictions predictive accuracy =(TP+TN)/(TP+TN+FP+FN)
- 9. Data mining techniques Decision Tree Regression artificial neural networks cluster analysis Association rules
- 10. Tools & platforms for data mining simple or a sophisticated standalone or embedded open source or a commercial User interface Data formats
- 11. comparison of popular data mining platform
- 12. Data mining best practices Business understanding data understanding data preparation modeling model evaluation discrimination and rollout
- 13. CRISP-DM data mining cycle
- 14. Myths about Data Mining Myth #1 Data mining is about algorithms Myth #2 Data mining is about predictive accuracy Myth #3 Data mining requires a data warehouse Myth #4 Data mining requires large quantities of data Myth #5 Data mining technology expert
- 15. Data mining mistakes Mistake #1 Selecting the wrong problem for data mining Mistake #2 Buried under mountains of data without clear metadata Mistake #3 Disorganized data mining Mistake #4 Insufficient business knowledge Mistake #5 Incompatibility of data mining tools and datasets Mistake #6 Looking only at aggregated results and not at individual records/predictions Mistake #7 Not measuring your results differently from the way you are sponsor measures them