It has the same accuracy with LDA in the case of all features, and it performs slightly better when only using petals. Create a histogram for each feature: Note that for both petal_length and petal_width, there seems to be a group of data points that have smaller values than the others, suggesting that there might be different groups in this data. Its body is short and the maximum is around 3 meters. The following types of snake species are feared because the toxins produced can kill their prey very quickly. petal_length and petal_width. In fact, most humans will kill the snake when they meet. Bodo is one of the most phenomenal pythons. Let’s build one using sklearn (documentation), with a maximum depth of 3, and we can check its accuracy on the test data: This decision tree predicts 98.3% of the test data correctly. How does history say that snakes are one of the most dangerous animals in the world. If we use multivariate Gaussian distribution to calculate the class conditional density instead of taking a product of univariate Gaussian distribution (used in Naive Bayes), we would then get a LDA model (read more here). Several options are available, including using kdeplot() to draw KDEs: sns. Please try enabling it if you encounter problems. Their maximum longevity is 13,8 years and normally people also called them scrub python. As an example, the popular dataset House Prices: Advanced Regression Techniques from Kaggle has about 80 features and more than 20% of them contain some level of missing data. Its habitat is now spread in Sumatra, Java, and Bali. Some features may not work without JavaScript. Although in general the public knows the types of pythons as non-venomous snakes, in fact at certain levels there can still be content in the saliva. The ball python (Python regius), also called royal python, is a python species native to West and Central Africa, where it lives in grasslands and shrublands. He has wrestled full grown bull apes and gorillas, lions, rhinos, crocodiles, pythons, sharks, tigers, man-size seahorses … We saw that the petal measurements are more helpful at classifying instances than the sepal ones. This python is called reticulated batik because it has a patterned skin like batik. Many people consider this snake quite docile and friendly, and happily welcome it near their home or farm because it eats rats and mice. 2. It is a popular model since it is relatively simple and easy to implement. It is NOT that the preparation stage is not important. All data are stored in a central database and there are tools available for extracting, inspecting, analyzing, and plotting of data, models, and results. This giant snake can prey on buffalo even human. Actually, Indonesia has many types of python, more than 12 types above. Its main characteristic is a reddish body color, resembling blood. Make learning your daily ritual. To visualize the decision boundary in 2D, we can use our LDA model with only petals and also plot the test data: Four test points are misclassified — three virginica and one versicolor. Both matplotlib and seaborn have great plotting tools then we can use for visualization. Furthermore, most models achieved a test accuracy of over 95%. Now we are ready to build some classifiers (woo-hoo!). Ball python-Wikipedia. The task is to classify iris species and find the most influential features. It’s quite long, ranging from 2.5 to 4 meters. Usually, we should also have a validation set, which is used to evaluate the performance of each classifier, fine-tune, and determine the best model. Its body that resembles a hose, scaly, slippery and able to attack with the blink of an eye with deadly poison makes humans rarely approach or even avoid this one animal. In Indonesia, this python is found on the islands of Sumatra, Sulawesi, Nusa Tenggara, to Maluku. See my code on Github for details. Hearing the name of this type of reptile does make a shudder. First, let’s look at a numerical summary of each attribute through describe: We can also check the class distribution using groupby and size: We can see that each class has the same number of instances.