Creating Your First Predictive Model In Python
3 likes8,691 views
This document outlines a step-by-step guide for creating a predictive model in Python, focusing on various aspects such as selecting the right modeling technique, preparing data, validating results, and implementing the model in production. Key topics include feature selection, training the model, and ensuring accuracy over time. It emphasizes the use of libraries like Scikit-learn and techniques like logistic regression, training/testing splits, and model serialization with pickle.
Creating Your First Predictive Model In Python
- 1. Creating Your First Predictive Model In Python
- 3. Information Everywhere Makes Panda sad and confused
- 4. Each New Thing You Learn Leads to another new thing to learn, and anotherâŠ
- 5. So Many Things 1. Which predictive modeling technique to use 2. How to get the data into a format for modeling 3. How to ensure the ârightâ data is being used 4. How to feed the data into the model 5. How to validate the model results 6. How to save the model to use in production 7. How to implement the model in production and apply it to new observations 8. How to save the new predictions 9. How to ensure, over time, that the model is correctly predicting outcomes 10.How to later update the model with new training data
- 7. Format The Data âą Pandas FTW! âą Use the map() function to convert any text to a number âą Fill in any missing values âą Split the data into features (the data) and targets (the outcome to predict) using .values on the DataFrame
- 8. Get The Right Data âą This is called âFeature selectionâ âą Univariate feature selection âą SelectKBest removes all but the k highest scoring features âą SelectPercentile removes all but a user-speciïŹed highest scoring percentage of features using common univariate statistical tests for each feature: false positive rate âą SelectFpr, false discovery rate SelectFdr, or family wise error SelectFwe. âą GenericUnivariateSelect allows to perform univariate feature selection with a conïŹgurable strategy. http://scikit-learn.org/stable/modules/feature_selection.html
- 9. Data => Model 1. Build the model http://scikit-learn.org/stable/modules/cross_validation.html from sklearn import linear_model logClassiïŹer = linear_model.LogisticRegression(C=1, random_state=111) 2. Train the model from sklearn import cross_validation X_train, X_test, y_train, y_test = cross_validation.train_test_split(the_data, the_targets, cv=12, test_size=0.20, random_state=111) logClassiïŹer.ïŹt(X_train, y_train)
- 10. Validation! 1. Accuracy Score http://scikit-learn.org/stable/modules/cross_validation.html from sklearn import metrics metrics.accuracy_score(y_test, predicted) 2. Confusion Matrix metrics.confusion_matrix(y_test, predicted)
- 11. Save the Model Pickle it! https://docs.python.org/3/library/pickle.html import pickle model_ïŹle = "/lr_classiïŹer_09.29.15.dat" pickle.dump(logClassiïŹer, open(model_ïŹle, "wb")) Did it work? logClassiïŹer2 = pickle.load(open(model, "rb")) print(logClassiïŹer2)
- 12. Implement in Production âą Clean the data the same way you did for the model âą Feature mappings âą Column re-ordering âą Create a function that returns the prediction âą Deserialize the model from the ïŹle you created âą Feed the model the data in the same order âą Call .predict() and get your answer
- 13. Save Your Predictions As you would any other piece of data
- 14. Ensure Accuracy Over Time Employ your minion army, or get more creative
- 15. Update the Model Train it again, but with validated predictions