JPH0287225A - Condition deciding device - Google Patents

Abstract

PURPOSE:To discriminate whether the spoken contents of a range designated by time belong to a specified condition or not by retrieving a feature word to the spoken contents of the range to be designated by the time and adding an importance degree. CONSTITUTION:A retrieving device 14 retrieves the spoken contents outputted by a spoken contents extracting device 12, and when there is a word coincident with the feature word which is recorded to a feature word recorder 15, the importance degree corresponding to the feature word is read and outputted to an adder 16. The retrieving device 14 executes the retrieval until this processing reaches the whole spoken contents outputted by the spoken contents extracting device 12. The adder 16 successively adds the importance degree outputted by the retrieving device 14. When the processing of the retrieving device 14 is finished, an added result is outputted to a deciding device 18. Thus, the added result is compared with a decision reference registered in advance in the deciding device 18. When the added result is larger than a reference value, the spoken contents belong to the specified condition and when the added result is smaller, the spoken contents do not belong to the specified condition.

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention determines whether or not a specified range of utterances among utterances during a certain task belong to a certain specific situation. The present invention relates to a situation determination device.

(Conventional technology) As a method for evaluating the usability of computers, etc.
There is a method in which the content of utterances during work is recorded on a magnetic disk or the like, and the content of the utterances is analyzed later.

(Problems to be Solved by the Invention) In general, the content of utterances during work includes utterances in various situations. However, if you want to analyze the content of utterances in a specific situation (for example, a situation in which a worker is in trouble), it is necessary to carefully read the content of the utterances printed on paper in order to classify the situation in advance. was there. (Means for Solving the Problem) The means provided by the present invention for extracting the content of utterances uttered in a certain specific situation, which is necessary for estimating the work situation, extracts the content of utterances during work and the content of each utterance. an utterance content recording device that records the time as a character string, a range specification device that specifies utterance content that does not determine the work status by time, and a range specification device that records the utterance content specified by the range specification device from the utterance content recording device. An utterance content extraction device for extraction, a feature-side recording device that records words that are frequently uttered in a specific situation as feature words, and a numerical value that records the importance of the feature words. Searching the utterance content extracted by the utterance content extraction device for words that match the feature words recorded in the importance recording device and the feature-side recording device, and further determining the importance level corresponding to the searched feature word. A search device that reads and outputs from the importance level recording device, an addition device that adds up the importance levels output by the search device, and a comparison between the addition result of the addition device and a reference value to specify the range. and a determination device that determines whether or not the utterance content within the range specified by the device belongs to the specific situation.

(Effect) Generally, the content of utterances in a specific situation during a certain task (
For example, in the content of utterances when a worker is "in trouble" while working on a computer, there are characteristic words that are uttered particularly frequently in that situation. Therefore, when a certain range of utterance content is given, by checking the number of characteristic words contained within it, it is possible to estimate whether the utterance content belongs to a specific situation. . However, among the characteristic words, there are words that are uttered relatively often even in situations other than specific situations, and words that are rarely uttered outside of specific situations. therefore,
If you add an importance level to each feature word that corresponds to the probability that it will be uttered in a certain situation, add this importance level during a single search, and compare the size of the addition results, you can get more accurate results. can be determined.

(Embodiment) FIG. 1 is a block diagram showing an embodiment of the present invention.
The contents of the worker's utterances during the work and the time at which each utterance was made are recorded, as shown by reference numeral 21 in FIG. The range specifying device 13 is an utterance content recording device that specifies a range in which the situation in the utterance content is to be determined by specifying a start time and an end time. 12 extracts the utterance content within the time range specified by the range specifying device 13 from the utterance content recording device 11. The operation of the present invention described below is such that the characteristic word recording device 15, which is characterized by the extracted utterance content, records words that are uttered with a particularly high frequency in a particular situation as special rj! 1. In the 0 importance recording device 17 recorded as a word, each ′jI corresponds to each feature word recorded in the feature side recording device 15.
The L requirement is recorded numerically. The degree of importance is, for example, the probability that the characteristic word is uttered in a specific situation in other utterance data, expressed as a relative value. Therefore, it can be said that the higher the importance, the higher the probability that the feature word will be uttered in the keyword specific situation. The search device 14 searches the utterance content outputted by the utterance content extraction device 12, and if there is a word that matches the feature word recorded in the feature word recording device': It15, it determines the importance level corresponding to the feature word. The data is read from the degree recording device 17 and output to the addition device 16. The retrieval device 14 performs the above processing until the entire utterance content outputted by the utterance content extraction device 12 is covered.The addition device 16 adds up the importance outputted by the retrieval device 14 one by one until the processing of the retrieval device 14 is completed. When the addition is completed, the addition result is output to the determination device 18.

In the determination device 18, numerical values serving as determination criteria are registered in advance for the addition results of the addition device.

If the addition result is larger than this reference value, it is determined that the utterance content specified by the range specifying device 13 belongs to a specific situation, and if it is smaller, it is determined that it does not belong.

The entire operation will be explained using the conceptual diagram of FIG. 2 as an example. In FIG. 2, an example of the utterance content recorded in the utterance content recording device 11 is indicated by reference numeral 21. If the range specifying device 13 specifies, for example, 150'' and ``190'' as the start time and end time for this utterance content, the utterance content extracting device 12 will extract the utterances within the range shown by reference numeral 22 in FIG. Extract the content. Now, the feature word recording device 15
2, and the importance level recording device 17 records the importance level as shown in 24 in FIG. 2 corresponding to each characteristic word. do.

The search device 14 searches the utterance content 22 for a word that matches the feature word 23, and if there is a matching word, reads the importance corresponding to the feature word from the importance recording device 17. In the example in Figure 2, "I°laira", "No", and "I don't understand" match the characteristic words, so their importance level is "6".
tl 411. “8” is read. Adding device 16
When the above processing is completed, the importance level read by the search device 14 is added to the entire speech content extracted by the speech content extraction device 12, and the result is output to the determination device. In the example of FIG. 2, 6+4+8=18 is output to the determination device. Based on the magnitude of this numerical value, it is determined whether the extracted utterance content belongs to a specific situation. Now, for example, if 15 is registered as the reference value, the addition result is larger than the reference value, so it is determined that the utterance content 22 in FIG. 2 belongs to a specific situation.

(Effects of the Invention) As is clear from the above explanation, according to the present invention, by searching for characteristic words and adding their importance to the utterance content within the range specified by the time, It is possible to determine whether a range of utterance content belongs to a certain specific situation or not.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the invention, and FIG. 2 is a conceptual diagram for explaining the invention in detail. 11... Utterance content recording device, 12... Utterance content extraction device, 13... Range specification device, 14... Search device,
15... Characteristic word recording device, 16... Addition device, 17
...Importance recording device, 18... Judgment device.

Claims (1)

[Claims] an utterance recording device that records the utterances during work and the time each utterance was made as a character string; a range specifying device that specifies the utterance content by time for which the work status is to be determined; an utterance content extraction device that extracts utterance content specified by a range specification device; a feature word recording device that records words that are frequently uttered in a specific situation as feature words; An importance level recording device that numerically records the importance level and words that match the characteristic words recorded in the characteristic word recording device are searched from the utterance content extracted by the utterance content extraction device, and further searched. a search device that reads and outputs the importance level corresponding to the feature word that has been searched from the importance level recording device; an addition device that adds up the importance levels output by the search device; and a size of the addition result of the addition device and a reference value. a determination device that determines whether or not the utterance content within the range designated by the range designation device belongs to the specific situation by comparing the range designation device.