YSKRecognitionHypothesis class

#import <YSKRecognitionHypothesis.h>

Object representing a single recognition hypothesis.

Each hypothesis has a confidence score — a number between 0 and 1 reflecting the level of confidence in its "correctness".

A hypothesis has two representations:
  1. A string containing the hypothesis text. This can be considered the primary representation. Numbers in this text are usually written as digits, and punctuation and abbreviations are included.

  2. An array of words or phrases, each of which has a confidence score. This score can be used to determine poorly recognized fragments.

Note.

The confidence score is provided solely for informational purposes. Its calculation formula may be adjusted with time.

Instance methods

Properties

confidence

The confidence score is a fractional number between 0 and 1 estimating the measure of confidence in the recognition result (1 indicates the highest level of confidence).

normalized

This can be considered the primary representation. Numbers in this text are usually written as digits, and punctuation and abbreviations are included.

words

This score can be used to determine poorly recognized fragments.

Method descriptions

-description

- (NSString *) description

-initWithWords:normalizedText:confidence:

- (instancetype) initWithWords:(NSArray *) words
normalizedText:(NSString *) normalized
confidence:(float) confidence

Property descriptions

confidence

(assign, readonly) float confidence

Confidence score of the recognition hypothesis.

The confidence score is a fractional number between 0 and 1 estimating the measure of confidence in the recognition result (1 indicates the highest level of confidence).

normalized

(copy, readonly) NSString * normalized

A string containing the hypothesis text.

This can be considered the primary representation. Numbers in this text are usually written as digits, and punctuation and abbreviations are included.

words

(copy, readonly) NSArray * words

An array of words, each with a confidence score.

This score can be used to determine poorly recognized fragments.