def main(): data = dataset_block(load_cancer_dataset(), withOnes=False) data_train, data_test = split(data) reg_const = optimize_regularization(data) theta = svm.train(data_train, c=reg_const) stats = svm.test(data_test, theta) print("precision:%6.2f\nrecall:%6.2f\nerror:%6.2f\nf1-score:%6.2f\n" % (stats.precision(), stats.recall(), stats.error(), stats.f_score())) print("regularization constant used: %6.2f\n" % reg_const)
def optimize_regularization(data): reg_best, f1_best = 0, 0 data_train, data_test = split(data) for d in range(-10, 40): reg_current = 0.5 ** d theta = svm.train(data_train, c=reg_current) stats = svm.test(data_test, theta) f1_current = stats.f_score() if f1_best < f1_current: reg_best, f1_best = reg_current, f1_current return reg_best
def optimize_regularization(data): reg_best, f1_best = 0, 0 data_train, data_test = split(data) for d in range(-10, 40): reg_current = 0.5**d theta = svm.train(data_train, c=reg_current) stats = svm.test(data_test, theta) f1_current = stats.f_score() if f1_best < f1_current: reg_best, f1_best = reg_current, f1_current return reg_best
from flask import redirect import download_corpora import NLProcessor as nlp #FIX ABOVE 2 LILNES IM NOT SURE WHAT TO DO from ml import svm import otherAPIs import scraper import searchFunction import stream import boto3 import lxml.html import requests from requests import get from goose import Goose svm.train() def getSuggestions(query): url = 'https://api.cognitive.microsoft.com/bing/v5.0/suggestions/?q=' + query headers = {'Ocp-Apim-Subscription-Key': '854e8088bb8347418e6f934b996487af'} r = requests.get(url, headers=headers) results = [] suggestions = r.json()['suggestionGroups'] max = 3 for suggestion in suggestions: s = suggestion['searchSuggestions'] for term in s: