Signal judgment method, signal judgment apparatus, program, and signal judgment system
Provided is a high-accuracy signal judgment method for noise or vibrations of a living environment. Provided is a signal judgment method associated with noise or vibrations of a living environment. The signal judgment method includes: inputting, into a model equation of a to-be-measured/analyzed target, unknown data; and obtaining an output value as a probability value of whether the unknown data is a to-be-measured/analyzed target. The model equation of the to-be-measured/analyzed target is created with the use of measured already-known data. The model equation is created by prediction model equation calculation means. Incidentally, a logistic regression equation can be used as the prediction model equation.
1. A signal judgment method that is a sound noise or vibrations judgment method of determining by a computer whether data is a signal judgment target, comprising:
inputting, into a model of a signal judgment target associated with the sound noise or vibrations, unknown data for which a type of the sound noise or vibrations is unknown; and,
obtaining an output value of the model of the signal judgment target as a probability value of whether the unknown data is a signal judgment target.
2. The signal judgment method according to claim 1 , wherein
the model of the signal judgment target is created with the use of measured already-known data for which a type of the sound noise or vibrations is already known.
3. The signal judgment method according to claim 2 , wherein:
the model of the signal judgment target is created by prediction model equation calculator with the use of the already-known data;
unknown data is input into the created model of the signal judgment target associated with the sound noise or vibrations; and
an output value calculated by an output value calculator that uses model of the signal judgment target is obtained as a probability value which is a prediction probability that the unknown data is a signal judgment target.
4. The signal judgment method according to claim 3 , wherein
the signal judgment target is a model of a signal judgment target that uses data concerning a sound noise or vibrations of a living environment.
5. The signal judgment method according to claim 4 , wherein:
a logistic regression equation is used for a prediction model equation associated with the model of the signal judgment target;
the logistic regression equation uses a prediction factor associated with the model of the signal judgment target as an explanatory variable;
the logistic regression equation uses the fact of whether the unknown data is the signal judgment target as an response variable; and
the probability value which is a prediction probability is calculated by adapting and inputting the unknown data to the logistic regression equation.
6. The signal judgment method according to claim 5 , wherein
some of the explanatory variables are adapted to prediction model equation after being classified into categories.
7. The signal judgment method according to claim 6 , wherein
the adaptation to the prediction model equation is made with the use of a degree of similarity between a plurality of items of the unknown data and sample data of a signal that becomes a signal judgment target, and the probability value which is a prediction probability is calculated.
8. The signal judgment method according to claim 7 , wherein
the adaptation to the prediction model equation is made with the use of the degree of dissimilarity between a plurality of items of the unknown data and sample data of a signal that is not a signal judgment target, and the probability value which is a prediction probability is calculated.
9. The signal judgment method according to claim 8 , wherein
data of a type that has a significant impact to an evaluation amount in case of being mistakenly determined to be a noise on a prediction model equation of the signal judgment target is used as sample data of a signal that is not a signal judgment target.
10. The signal judgment method according to claim 7 , wherein
the sample data is created with the use of a statistical method from sampled time-series data of a predetermined period of time.
11. The signal judgment method according to claim 5 , wherein
a model equation of a signal judgment target is created by calculating a coefficient of an explanatory variable of the logistic regression equation with the use of the already-known data.
12. The signal judgment method according to claim 5 , wherein
a process of carrying out a test of a prediction model equation of the prediction model or explanatory variable and a process of carrying out an examination or test of validity are further performed.
13. The signal judgment method according to claim 5 , wherein
the already-known data is accumulated, and the prediction model equation is created again on the basis of the accumulated data.
14. The signal judgment method according to claim 3 , wherein
a plurality of signal judgment targets is changed, the unknown data is applied to the prediction model equation for the changed signal judgment target, and a probability value which is a prediction probability that the unknown data is the changed signal judgment target is obtained.
15. The signal judgment method according to claim 3 , wherein
a plurality of the prediction model equations are provided, and, if one portion of the unknown data is not acquired, a prediction model equation that does not apply the unknown data is used among a plurality of the prediction models equations.
16. A non-transitory computer readable medium storing a program which causes a processor to perform the signal judgment method according to claim 1 when executed by the processor.
17. A signal judgment apparatus, comprising:
storage for storing a prediction equation of a model of a signal judgment target associated with a sound noise or vibrations, already-known data for which a type of the sound noise or vibrations is already known, unknown data for which a type of the sound noise or vibrations is already unknown, sample data of a signal that is to be judged, and sample data of a signal that is not to be judged;
prediction model equation calculator for creating a prediction model equation by calculating a coefficient of an explanatory variable with the use of a degree of similarity or dissimilarity between the already-known data and sample data of a signal that becomes a signal judgment target and/or sample data of a signal that is not a signal judgment target;
an input unit configured to input an input value to the storage;
an output value calculator configured to apply and calculate each value of the unknown data to the prediction model equation of a model of the signal judgment target;
an output unit configured to output the output value; and
a judgment unit configured to measure and/or analyze a type of the unknown data by using the output value.
18. The signal judgment apparatus according to claim 17 , wherein:
the storage is a storage in which a plurality of the prediction model equations is stored when the explanatory variable is added;
the signal judgment apparatus further includes a data integrator configured to integrate output values of a plurality of the prediction model equations; and
the output value calculator applies data concerning the explanatory variable to a plurality of the prediction model equations by the integrator for calculation.
19. A signal judgment apparatus, comprising:
the signal judgment apparatus claimed in claim 17 or 18 ; and
a sensor that transmits the unknown data to the input unit.