About this repository
This repository is a Flask App that has a machine learning model merged with it and deployed on heroku.
This predicts famous iris species among Versicolor, Setosa and Virginica
Here is picture of thosed species
ML model used
In this repository I have used Linear Support Vector Classsification model pre-built in sklearn for predicting classes.
LinearSVC works well for datasets having less data and less atrributes to train the model. Dataset used has only 150 data samples overall so
LinearSVC performed well with accuracy of 93% and predicted only 3 test data wrong.
Predicted classes visualization
Scatter plot between attributes and categorized using color based on predictions
Petal width VS Petal length
Petal width VS Petal length
More visualization can be found in jupyter notebook in this repository itself named: Flowers Identification using svm.ipynb
Libraries used
- Numpy
- Pandas
- Seaborn
- Matplotlib
- Sklearn