![About Electricity Data Set
•Electrical Grid data set in which we have different attributes for examine the stability of
the system.
• We will examine the response of the system stability depends on 10,000 observations
and 13 attributes, 1 classes attribute (stabf). Attributes are given in dataset as;
● tau[x] : Reaction time of participant (real from the range [0.5,10]s).
● p[x] : Nominal power consumed(negative)/produced(positive)(real).
● g[x] : Coefficient (gamma) proportional to price elasticity (real from the range
[0.05,1]s^-1).
● stab: The maximal real part of the characteristic equation root (if positive - the
system is linearly unstable)(real)
● stabf: The stability label of the system (stable/unstable)](https://image.slidesharecdn.com/mlprojectpresentation3-190804204130/85/Machine-Learning-project-presentation-3-320.jpg)
Machine Learning project presentation
- 1. Machine learning implementations on different datasets using python Submitted by: Oanh Doan Tess Mundadan Ramandeep Kaur Bagri
- 2. Problem formulation ● Electricity data set ● This is classification problem of electrical grid stability into two classes; ● 1.Stable ● 2.Unstable ● Olivetti Dataset ● Recognizing face with highest accuracy using machine learning algorithms ● Comparison of their accuracies to select the most fitted face recognition algorithm.
- 3. About Electricity Data Set •Electrical Grid data set in which we have different attributes for examine the stability of the system. • We will examine the response of the system stability depends on 10,000 observations and 13 attributes, 1 classes attribute (stabf). Attributes are given in dataset as; ● tau[x] : Reaction time of participant (real from the range [0.5,10]s). ● p[x] : Nominal power consumed(negative)/produced(positive)(real). ● g[x] : Coefficient (gamma) proportional to price elasticity (real from the range [0.05,1]s^-1). ● stab: The maximal real part of the characteristic equation root (if positive - the system is linearly unstable)(real) ● stabf: The stability label of the system (stable/unstable)
- 4. Head of the data set
- 6. Head of the data set
- 8. Classification using PCA Method 90:10 80:20 75:25 70:30 Confusion Matrix Naive bayes 0.706 0.6985 0.6932 0.6997 KNN (N=12) 0.686 0.6735 0.666 0.6723 SVM 0.694 0.689 0.6848 0.6897 Decision Tree 0.618 0.622 0.6352 0.6216 Random Forest 0.661 0.6655 0.6612 0.6703
- 9. Classification using without PCA,use standardization Method 90:10 80:20 75:25 70:30 Confusion matrix Naive Bayes 0.831 0.832 0.833 0.83466 KNN (N=12) 0.906 0.9102 0.9108 0.909 SVM 0.81 0.812 0.8108 0.8123 Decision Tree 0.879 0.8715 0.869 0.85766 Random Forest (n_estimators=6 0) 0.923 0.9215 0.916 0.9186
- 10. Classification by ratio 95:5 Method Accuracy Confusion Matrix Naive Bayes 0.832 KNN 0.92 SVM 0.92 Decision Tree 0.878 Random Forest 0.826
- 11. Selecting different features and results Naive Bayes: Accuracy - 0.982 KNN: Accuracy - 0.964 Decision Tree Accuracy - 1.0 Random Forest: Accuracy - 1.0 SVM Accuracy - 99.2 Test: 0.05
- 12. Best algorithm fit for electricity simulated data Set for classification Random Forest( Using n_estimators = 60 ): ● Test dataset: 90:10 ● Accuracy = 0.923 ● Confusion Matrix: ● Although the KNN and SVM has given good accuracy at ratio of 95:5, but it is only on test dataset – need further evaluation on training dataset. ● By selecting different features the accuracy observed is 1, but the model is overfitted – Need prepossing and proper selection of features to train the models to improve accuracy.
- 13. Step 1: Load and explore the data
- 14. Step 2: Visualize the data Olivetti Data set
- 16. Reshape data
- 18. Accuracy Metrics Comparison without PCA Method 90:10 80:20 70:30 Naive_Bayes 92.5 87.50 73.33 KNN 90.0 91.25 90.00 Random Forest 90.0 93.75 93.33 SVM 95.0 96.25 96.67 Logistic Regression 95.0 97.50 97.50
- 19. Dimensionality Reduction using PCA
- 20. Mean Face of the Samples
- 21. Eigen Faces
- 22. Metrics with different Principle Components Method PC = 90 PC =103 PC = 200 Naive Bayes 76.67 77.50 69.17 KNN 90.83 90.83 90.00 Random Forest 93.33 92.50 91.67 Logistic Regression 96.67 98.33 98.33 SVM 96.67 96.67 96.67
- 23. Accuracy Metrics Comparison with PCA METHOD 90:10 80:20 70:30 Without PCA With PCA Without PCA With PCA Without PCA With PCA Naive Bayes 92.5 92.5 87.50 90.0 73.33 77.50 KNN 90.0 90.0 91.25 92.5 90.00 90.83 Random Forest 90.0 100.0 93.75 95.0 93.33 92.50 Logistic Regression 95.0 97.5 96.25 97.5 97.50 98.33 SVM 95.0 95.0 97.5 97.5 96.67 96.67
- 24. Best fit for Olivetti data set Random Forest : PCA = 103 Test dataset = 90:10 Accuracy = 100
- 25. Conclusion Use of PCA improved the accuracy metrics. Project Enhancement Idea : Identify smile and no Smile face/ Identify male and female
- 26. Reference ● https://www.kaggle.com/serkanpeldek/face-recognition-on-olivetti-dataset ● https://colab.research.google.com/notebooks/welcome.ipynb ● https://www.youtube.com/watch?v=PitcORQSjNM ● https://www.youtube.com/watch?v=_VTtrSDHPwU&t=86s ● https://www.kaggle.com/elikplim/eergy-efficiency-dataset/kernels ● Class notes: Introduction to data mining and machine learning
- 27. Thank you for your attention
Editor's Notes
- #15: Data is collected by taking the pictures of 40 different human personalities in each column of the dataset. Each row is the sample of one human personality with different face emojis to train the dataset. Dimensions: 10:400
- #19: No N/As Machine learning models can work on vectors. Since the image data is in the matrix form, it must be converted to a vector. Different testing datasets are being used: 0.1,0.2,0.3.
- #24: SVM is not affected by PCA 100 is due to small and clean data set Increasing the percentage of training set improves accuracy