Python implementation of a simple Bayesian classifier

1 #-*-coding:utf-8-*-2 " "3 >>> C = Classy ()4 >>> c.train ([' CPU ', ' RAM ', ' ALU ', ' io ', ' bridge ', ' disk '], ' architecture ')5 True6 >>> C.train ([' Monitor ', ' mouse ', ' keyboard ', ' microphone ', ' headphones '], ' input_devices ')7 True8 >>> c.train ([' Desk ', ' chair ', ' cabinet ', ' lamp '], ' office furniture ')9 TrueTen >>> my_office = [' CPU ', ' monitor ', ' mouse ', ' chair '] One >>> c.classify (my_office) A (' input_devices ', -1.0986122886681098) - ... - >>> C = Classy () the >>> c.train ([' CPU ', ' RAM ', ' ALU ', ' io ', ' bridge ', ' disk '], ' architecture ') - True - >>> C.train ([' Monitor ', ' mouse ', ' keyboard ', ' microphone ', ' headphones '], ' input_devices ') - True + >>> c.train ([' Desk ', ' chair ', ' cabinet ', ' lamp '], ' office furniture ') - True + >>> my_office = [' CPU ', ' monitor ', ' mouse ', ' chair '] A >>> c.classify (my_office) at (' input_devices ', -1.0986122886681098) - ... - " " -  -  fromCollectionsImportCounter - ImportMath in  - classclassifiernottrainedexception (Exception): to      +     def __str__(self): -         return "Classifier is not trained." the  * classClassy (object): $     Panax Notoginseng     def __init__(self): -Self.term_count_store = {} theSelf.data = { +             'Class_term_count': {}, A             'beta_priors': {}, the             'Class_doc_count': {}, +         } -Self.total_term_count =0 $Self.total_doc_count =0 $          -     deftrain (self, Document_source, class_id): -      the         " " - Trains the classifier.Wuyi          the         " " -Count =Counter (Document_source) Wu         Try: - self.term_count_store[class_id] About         exceptKeyerror: $SELF.TERM_COUNT_STORE[CLASS_ID] = {} -          forTerminchCount: -             Try: -Self.term_count_store[class_id][term] + =Count[term] A             exceptKeyerror: +Self.term_count_store[class_id][term] =Count[term] the         Try: -self.data['Class_term_count'][CLASS_ID] + = Document_source.__len__() $         exceptKeyerror: theself.data['Class_term_count'][CLASS_ID] = Document_source.__len__() the         Try: theself.data['Class_doc_count'][CLASS_ID] + = 1 the         exceptKeyerror: -self.data['Class_doc_count'][CLASS_ID] = 1 inSelf.total_term_count + = Document_source.__len__() theSelf.total_doc_count + = 1 the self.compute_beta_priors () About         returnTrue the          the     defclassify (self, document_input): the         if  notSelf.total_doc_count:Raiseclassifiernottrainedexception () +          -Term_freq_matrix =Counter (document_input) theArg_max_matrix = []Bayi          forclass_idinchself.data['Class_doc_count']: thesummation =0 the              forTerminchDocument_input: -                 Try: -Conditional_probability = (Self.term_count_store[class_id][term] + 1) theConditional_probability = conditional_probability/(self.data['Class_term_count'][CLASS_ID] +self.total_doc_count) theSummation + = term_freq_matrix[term] *Math.log (conditional_probability) the                 exceptKeyerror: the                      Break -Arg_max = summation + self.data['beta_priors'][class_id] the Arg_max_matrix.insert (0, (class_id, Arg_max)) theArg_max_matrix.sort (key=LambdaX:x[1]) the         return(Arg_max_matrix[-1][0], arg_max_matrix[-1][1])94          the     defcompute_beta_priors (self): the         if  notSelf.total_doc_count:Raiseclassifiernottrainedexception () the         98          forclass_idinchself.data['Class_doc_count']: AboutTMP = self.data['Class_doc_count'][CLASS_ID]/Self.total_doc_count -self.data['beta_priors'][CLASS_ID] = Math.log (TMP)

Python implementation of a simple Bayesian classifier