Dataset Analysis using weka tools (pattern recognition)
Download as PPTX, PDF1 like795 views
This document analyzes and summarizes the performance of different classifiers (kNN, NBC, Decision Tree, oneR, Random Forest) on four datasets (Mushroom, Wine-Quality, Flags, ZOO). For each dataset, the classifiers are evaluated using 10-fold cross validation, and their accuracy, error rate, recall, precision and F-score are reported. The Random Forest classifier generally performed best, achieving an accuracy of 70.10% on the Wine-Quality dataset. References and links to the datasets are also provided.
Dataset Analysis using weka tools (pattern recognition)
- 1. Dataset Analysis Presented By Nazmul Hyder ID : 011 131 085 Section : SB
- 2. Contents ❑ Dataset Name ❑ Classifiers ❑ Dataset Description ❑ Dataset Analysis ❑ Graphical representation. ❑ References
- 3. Datasets Name ❏ Mushroom. ❏ Wine-Quality. ❏ Flags. ❏ ZOO.
- 5. Dataset Description Dataset name No of instances No of attributes Attribute type Class value Data denoted Donor Mushroom 8124 22 nominal 2 1987 Jeff Schlimmer Wine-Quality 1599 12 numeric 6 (nominal) 2009 Paulo Cortez, Antonio Cerdeira, Fernando Almeida Flags 194 30 nominal 194 (nominal) 1990 Richard S. Forsyth ZOO 101 17 nominal 8 (nominal) 1990 Richard S. Forsyth
- 6. Dataset Analysis: Mushroom-Cross validation(10 folds) Classifier Accuracy Error Rate Recall Precision F-score kNN (k=3%) 59.6135% 40.3865% 0.596 0.576 0.583 NBC 64.5126% 35.4874% 0.645 0.769 0.665 j4.8 61.9645% 38.0355% 0.620 0.629 0.623 oneR 57.9025% 42.0975% 0.579 0.411 0.469 Random Forest 47.3043% 52.6957% 0.473 0.476 0.474
- 7. Dataset Analysis (con.) Wine-Quality-Cross validation(10 folds) Classifier Accuracy Error Rate Recall Precision F-score kNN (k=3%) 57.7236% 42.2764% 0.577 0.542 0.553 NBC 55.0344% 44.9656% 0.550 0.554 0.550 j4.8 61.4759% 38.5241% 0.615 0.612 0.613 oneR 54.6592% 45.3408% 0.547 0.496 0.511 Random Forest 70.1063% 29.8337% 0.701 0.679 0.684
- 8. Flags - Cross validation(10 folds) Classifier Accuracy Error Rate Recall Precision F-score kNN (k=3%) 59.2789% 40.7216% 0.593 0.553 0.550 NBC 55.1546% 44.8454% 0.552 0.571 0.542 j4.8 59.2784% 40.7216% 0.593 0.570 0.576 oneR 4.6392% 95.3608% 0.046 0.002 0.004 Random Forest 61.3402% 38.6598% 0.613 0.545 0.572 Dataset Analysis (con.)
- 9. ZOO - Cross validation(10 folds) Classifier Accuracy Error Rate Recall Precision F-score kNN (k=3%) 94.1176% 5.8824% 0.941 0.935 0.931 NBC 95.098% 4.902% 0.951 0.953 0.950 j4.8 92.1569% 7.8431% 0.922 0.916 0.915 oneR 2.9412% 97.0588% 0.029 0.039 0.026 Random Forest 92.1569% 7.8431% 0.922 0.874 0.896 Dataset Analysis (con.)
- 11. References : Quick Links : Mushroom:https://archive.ics.uci.edu/ml/datasets/mushroom Wine Quality:https://archive.ics.uci.edu/ml/datasets/wine+quality Flags : https://archive.ics.uci.edu/ml/datasets/Flags ZOO: http://archive.ics.uci.edu/ml/datasets/Zoo URL : http://archive.ics.uci.edu/ml/datasets.html
- 12. Thank You