Created
January 18, 2017 14:15
-
-
Save BarnabasMarkus/d718f2dde1f23dda4a295ee86fd07f34 to your computer and use it in GitHub Desktop.
naive-bayes classifier
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| #!/usr/bin/env python3 | |
| # N A I V E B A Y E S C L A S S I F I E R | |
| # Project Naive-Bayes Classifier | |
| # Author Barnabas Markus | |
| # Email barnabasmarkus@gmail.com | |
| # Date 18.01.2017 | |
| # Python 3.6 | |
| # License MIT | |
| from collections import Counter, defaultdict | |
| from operator import mul | |
| from functools import reduce | |
| class NaiveBayes: | |
| def __init__(self): | |
| self.words_cats = defaultdict(lambda: defaultdict(int)) | |
| self.cats = defaultdict(int) | |
| self.thresholds = {} | |
| @staticmethod | |
| def get_words(doc): | |
| """Return list of words from str""" | |
| return [word.lower().strip() for word in doc.split() if len(word) > 2] | |
| def train(self, doc, cat): | |
| """Train the classifier with document-category pairs""" | |
| words = self.get_words(doc) | |
| for word in words: | |
| self.words_cats[word][cat] += 1 | |
| self.cats[cat] += 1 | |
| def word_probability(self, word, cat): | |
| """Return probability of a word belongs to a given category""" | |
| if word not in self.words_cats: | |
| return 0 | |
| return self.words_cats[word][cat] / sum(self.words_cats[word].values()) | |
| def weighted_word_probability(self, word, cat, weight=1.0, ap=0.5): | |
| # ap: asssumed_probability | |
| # weight: the weight of assumed_probability | |
| probability = self.word_probability(word, cat) | |
| totals = sum(self.words_cats[word].values()) | |
| return ((weight * ap) + (totals * probability)) / (weight + totals) | |
| def doc_probability(self, doc, cat): | |
| """Return probability of a doc belongs to a given category""" | |
| words = self.get_words(doc) | |
| weighted_probabilites = [self.weighted_word_probability(word, cat) | |
| for word in words] | |
| return reduce(mul, weighted_probabilites, 1) | |
| def cat_probability(self, cat): | |
| """Return probability of a category compared to all categories""" | |
| return self.cats[cat] / sum(self.cats.values()) | |
| def probability(self, doc, cat): | |
| cat_probability = self.cat_probability(cat) | |
| doc_probability = self.doc_probability(doc, cat) | |
| return cat_probability * doc_probability | |
| def set_threshold(self, cat, threshold): | |
| """Set thredhold for category""" | |
| self.thresholds[cat] = threshold | |
| def get_threshold(self, cat): | |
| """Get threshold of category""" | |
| if cat not in self.thresholds: | |
| return 1.0 | |
| return self.thresholds[cat] | |
| def classify(self, doc, default=None): | |
| """Classifing a document""" | |
| results = {} | |
| for cat in self.cats: | |
| results[cat] = self.probability(doc, cat) | |
| # best_1, best_2 = Counter(results).most_common(2) | |
| (cat_1, score_1), (_, score_2) = Counter(results).most_common(2) | |
| threshold = self.get_threshold(cat_1) | |
| if score_1 < score_2 * threshold: | |
| return default | |
| return cat_1 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment