Phase5-withRandom

The fifth phase of our project. In this phase we are hoping to create a final classifier of PDF files. The final classifier will be based on the three previous machines in our project: image, text, and features. The process of the final machine will be the following:

• install: sudo pip3 install xgboost
• Extract all data needed for the three base machines (image, text, features) - this is done using classes, imported into as.py.
• Create base vectors for every sample (image, text, features)
• Run every base machine on the samples, and return the calssification of the sample by every machine.
• Create a vector for the boost algorithm from the base machines classifications for every sample.
• Run boost algorithm with RF on sample boost vectors.
• Return boost algorithm accuracy.

all_random.png:

• AdaBoostClassifier: 8171 - true, 568 - false, accuracy - 93.5%.
• AdaBoostRegressor: 8414 - true, 325 - false, accuracy - 96.28%.
• XGBClassifier: 8180 - true, 559- false, accuracy - 93.6%.
• XGBRegressor: 8216 - true, 523 - false, accuracy - 94.01%.
• Random Forest Classifier: 7868 - true, 871 - false, accuracy - 90.03%.
• Random Forest Regressor: 8216 - true, 523 - false, accuracy - 94.01%.