###1.3rd party Dependency numpy pandas sklearn
###2.Modules There are 4 modules in this project: form10DB, DataGenerator, forecaster, and utils
2.1 form10DB form10DB is used to download and manage the 10 forms. It includes functions to download 10 forms for all symbols from 1993 - present. Form10Manager class is the workhorse to download data and populate them into a structured database.
2.2 DataGenerator DataGenerator module contains functions to generate structured dataset.
2.3 forecaster forecaster module contains classes and functions to train classifiers and make predictions.
2.4 utils utils contain usefule helper functions to support other modules.
###3.How to use In this demo project, the database and dataset have already been generated. So there is no need to use the module form10DB and DataGenerator. These two modulles are only needed when you want to repopulat the edgar database on your own machine.
The only module used in the demo project is forecaster. For demo, please change the working directory to edgar directory:
import forecaster learner = forecaster.Forecaster() ####read in training data for learner learner.readTrainFile('Dictionary.3label.Industrials','3label.Industrials.train.key') #####read in development data learner.readDevFile('3label.Industrials.dev.key') #####read in testing data learner.readTestFile('3label.Industrials.test.key') #####Training classifier learner.classifier.fit(learner.trainFeature,learner.trainLabel) #####Testing on development set learner.classifier.score(learner.devFeature,learner.devLabel) #####Get the accuracy on testing set learner.classifier.score(learner.testFeature,learner.testLabel)
###4.Documentation Please refer to Documentation.html for detailed function API and source code