JP3077820B2 - Inappropriate sentence detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for detecting an inappropriate portion of a sentence, which automatically detects an inappropriate portion of a given sentence.

[0002] The improper part mentioned here is a typographical error.
Omissions, grammatical errors, inconsistent notation (or unacceptable notation), confusing expressions, etc.

[0003]

2. Description of the Related Art A conventional apparatus for detecting an inappropriate portion of a sentence of this type is disclosed, for example, in Japanese Patent Application Laid-Open No. Sho 59-127140, "Automatic Error Inspection and Correction Device for Text", and Japanese Patent Application Laid-Open No. Sho 61-282,963, "Japanese Word Processor. And JP-A-62-209.
No. 662, “Japanese sentence proofreading device” and the like.

[0006] In these methods, the conditions of inappropriate locations are stored in advance, and locations matching the conditions are retrieved from given sentences and detected as inappropriate locations. Further, in order to perform the presentation of the correction candidate, not only the condition of the improper portion but also the method of presenting the correction candidate when the condition is met should be stored together.

[0005] For example, as a condition of an inappropriate portion, "when words having the same pronunciation but different notations are used in the same context (words before and after are the same), the one with a lower appearance frequency is detected." When there is (detection condition 1), the following example sentence "... is a device for detecting an inappropriate portion .......... The condition of an inappropriate portion is stored. In this case,... Are detected as inappropriate portions. ”(Example sentence 1) In the above, the“ suitable ”portion matches the detection condition 1 and is detected.

[0006] For the detection condition 1, the following correction method "present the one having a higher frequency of occurrence as a correction candidate" ...
If there is (correction method 1), in the example sentence 1, “appropriate” is presented as a correction candidate for “appropriateness”.

[0007] As another detection condition, for example, "detect a one-character word weakly connected to words before and after appearing in a kanji string" (detection condition 2) is as follows. In the example sentence "Detect unsuitable king portion" (Example sentence 2), the "king" portion is detected. Then, "replace the character with a character that is easily recognized by the character recognition device"
If there is the (correction method 2), “positive” is presented as a candidate for “king” (in this case, the relationship of characters that are easily erroneously recognized by the character recognition device is prepared in advance as a confusion matrix. ).

[0008]

A conventional apparatus for detecting an inappropriate portion of a sentence detects an inappropriate portion under the same conditions regardless of the input person and the input device using any input device. Correction candidates are presented.

[0009] However, the inappropriate factor that is likely to occur differs depending on the type of the text input device, and the degree of inappropriateness also differs. Furthermore, the tendency differs depending on the individual who inputs the sentence.

For example, the above-described detection condition 1 is a condition of an inappropriate portion that is likely to appear when a sentence is created as a kana-kanji conversion input device. However, when a sentence is input using a character recognition device, it is considered that improper portions rarely meet this condition.

An example of a more prominent detection condition is “detect a one-character word“ no ”immediately after a noun”.
(Detection condition 3) The correction method for this is “represent“ field ”with“ no ”” (correction method 3). An example sentence to which this applies is: "The effect of the invention is that the processing speed is improved."
(Example sentence 3) This corresponds to the current situation in which the particle "no" is erroneously converted to "no" during kana-kanji conversion. The detection condition 3 and the correction method 3 are preferably applied only when a sentence is input by a kana-kanji conversion input device. The tendency for good points to be detected increases.

Conversely, the above-mentioned correction method 2 is not appropriate for a sentence input by a kana-kanji conversion input device. If a Kana-Kanji conversion input device is used, there is a high possibility that the following correction method can present a correct correction candidate. "Present the same kanji" ... (correction method 2a) "Replace with the character corresponding to another adjacent key on the keyboard" ... (correction method 2b) Even when using the same kana-kanji conversion input device, there are individual differences in the points to be careful when inputting sentences. Therefore,
An input person who often makes an error like the detection condition 1 but hardly makes an error like the detection condition 3 can be considered.

A prominent example in which the detection condition differs depending on the individual (input person) is “detect a sentence having a length of 50 characters or more” (detection condition 4). Detect sentence "(detection condition 4a).

An object of the present invention is to provide an apparatus for detecting an inappropriate portion of a sentence in which a detection condition and a correction method can be changed according to a type of an input device and an input user.

[0015]

According to a first aspect of the present invention, there is provided an apparatus for detecting an inappropriate portion of a sentence which automatically detects an inappropriate portion of a given sentence. , The condition storage in the given sentence
Means based on the condition of the inappropriate location stored in the means.
Detecting means for detecting an inappropriate part in the sentence ,
Information on the type of the input device whether the input means of the given sentence is a kana-kanji conversion input device or a character recognition device
Or identification information that identifies the person who entered the information,
Input identification means for identifying at least one of the information, and condition changing means for changing the content stored in the condition storage means in response to the notification from the input identification means.

According to a second aspect of the present invention, there is provided an apparatus for detecting an improper part of a sentence for automatically detecting and correcting an improper part of a given sentence. And the condition in the given sentence
Based on the condition of the inappropriate part stored in the storage means
Detecting means for detecting an improper part in a given sentence; correcting means for applying a correction method to the part detected by the detecting means to generate a correction candidate; and kana-kanji input means for the given sentence any identification information for identifying a transformation type information input is either a character recognition device whether device input devices or input <br/>force's from these input information
Or at least one piece of information, and a condition changing means for changing contents stored in the condition storing means in response to a notification from the input identifying means.

[0017]

The first invention is an apparatus for detecting a mere improper part of a sentence, and does not include a correction function. Second
The present invention is an apparatus for detecting an inappropriate portion of a sentence including a correction function.

According to the first aspect of the present invention, the detection condition is changed in accordance with the type of the input device and the input user. That is, when the input device is a kana-kanji conversion input device, the above-described detection condition 3 is used as one of the detection conditions, but when the input device is a character recognition device, the detection condition 3 is excluded from the detection conditions to be used. Such changes can be made.
Alternatively, a change can be made such that the detection condition 4 and the detection condition 4a are interchanged according to the input person.

In the apparatus for detecting an inappropriate part of a sentence according to the second invention, not only the detection conditions but also the correction method are changed according to the type of the input device and the input person. That is, when the input device is a kana-kanji conversion input device, the above-described correction method 2a or 2b is associated with the detection condition 2 as a correction method,
When the input device is a character recognition device, it can be changed so as to correspond to the second correction method.

[0020]

Next, an embodiment of the present invention will be described.

FIG. 1 is a block diagram showing a configuration of an embodiment (referred to as a first embodiment) of an apparatus for detecting an inappropriate portion of a sentence according to the first invention. In FIG. 1 , a portion composed of a text storage unit 17, a condition storage unit 11, a detection unit 12, and a presentation unit 16 corresponds to a configuration portion of a conventional text incorrect position detection device. The text storage means 17 is a means for storing the input text, and is realized using an IC memory, a magnetic disk device, or the like. The condition storage means 11 is a means for storing the conditions of the inappropriate place such as the above-mentioned detection condition 1, detection condition 2, and so on, and is realized by using an IC memory, a magnetic disk device or the like. The detecting means 12 applies the condition storage means 1 to the sentence stored in the sentence storage means 17.
This is a means for searching for a portion that satisfies the detection condition stored in No. 1 and detecting the corresponding portion, and is realized by a CPU or the like. The presentation means 16 is a means for presenting the inappropriate portion detected by the detection means 12, and is realized by a display or a printer. As the presentation method by the presentation means 16, the relevant part in the text is displayed by marking such as inversion, color, underlining, shading, or the like, or only the relevant part (or its surroundings) is extracted and displayed. There are things to do.

The components newly required in the first embodiment include an input identifying means 13 and a condition changing means 14.
The input identification means 13 is a switch for notifying the type of the input device, the identification of the input person, and the like. The condition changing means 14 changes the contents stored in the condition storing means 11 in response to the notification from the input identifying means 13, and can be realized by a CPU and a memory. An example of a specific changing method in the condition changing unit 14 will be described. In the condition storage means 11, a detection condition independent of the type of the input device and the input person is written in advance. Such a detection condition includes, for example, the aforementioned detection condition 2. In the memory in the condition changing means 14,
The type of input device and the input user are classified in advance, and the detection conditions used in each case are stored in association with each other. For example, Kana-Kanji conversion input device (Detection condition 1) Input person A (Detection condition 4) Input person B (Detection condition 4a) When notified of the identification of a certain input device or a certain input person from the input identification means 13, the condition changing means 14 newly adds a detection condition corresponding thereto to the condition storage means 11 . If the input device / input person is changed from the input device / input person before the notification, the condition changing means 14 deletes the detection condition corresponding to the input device / input person from the condition storage means 11 and is newly notified. Write the detection conditions corresponding to the input device and the input user. For example, if the input person is A to B
When the condition is changed to, the detection condition stored in the condition storage means 11 is the case of the input person A: the detection condition 2 The detection condition 4 is deleted from the detection condition 4 and the detection condition 4a is added, and the detection condition is the input person B : Detection condition 2 Changes to detection condition 4a.

FIG. 2 is a block diagram showing a configuration of an embodiment (referred to as a second embodiment) of the apparatus for detecting an inappropriate portion of a sentence according to the second invention. In FIG. 2 , the sentence storage means 27,
The part constituted by the condition storage means 21, the detection means 22, the correction means 25, and the presentation means 26 corresponds to the components in a conventional apparatus for detecting an inappropriate portion of a sentence. Sentence storage means 27
And the detecting means 22 are the sentence storing means 1 in the first embodiment.
7 and the detecting means 12. The condition storage means 21 stores the condition of an inappropriate portion such as the above-described detection condition 1, detection condition 2, and so on, and a correction method (correction candidate generation method) such as correction method 1, correction method 2, and so on. And IC
This is realized using a memory, a magnetic disk device, or the like. Although the condition storage unit 11 in the first embodiment stores only the detection conditions, the condition storage unit 21 in the present embodiment stores not only the detection conditions but also the correction method. The correction unit 25 is a unit that generates a correction candidate by applying a corresponding correction method stored in the condition storage unit 21 to a location detected by the detection unit 22.
It is realized by a PU or the like. The presentation means 26 is the detection means 2
This is a means for presenting the incorrect location detected by the correction means 25 together with the correction candidate generated by the correction means 25, and is realized by a display or a printer.

The components newly required in the second embodiment include an input identifying means 23 and a condition changing means 24. The input identification means 23 is the same as the input identification means 13 in the first embodiment, and is a switch for notifying the type of the input device, the identification of the input person, and the like. The condition changing means 24 changes the contents stored in the condition storing means 21 in response to the notification from the input identifying means 23, and can be realized by a CPU and a memory. However, the condition changing means 1 of the first embodiment
No. 4 changes only the detection condition, but the condition changing means 24 of this embodiment changes not only the detection condition but also the correction method.

[0025]

As described above, according to the present invention, it is possible to change the search condition and the correction method according to the type of the input device, the input person, and the like.

[Brief description of the drawings]

FIG. 1 is a diagram showing a first example of an apparatus for detecting an incorrect portion of a sentence according to the present invention .
FIG. 3 is a block diagram illustrating a configuration of an example.

FIG. 2 shows a second embodiment of the apparatus for detecting an incorrect part of a sentence according to the present invention .
FIG. 3 is a block diagram illustrating a configuration of an example.

[Explanation of symbols]

 11,21 condition storage means 12,22 detection means 13,23 input identification means 14,24 condition change means 25 correction means 16,26 presentation means 17,27 sentence storage means

Claims (2)

(57) [Claims] 1. A automatically detect improper location device sentences to improper location of the sentences given, a condition storage means for storing the condition of the improper position, the condition storage in the given sentence Remembered by means
The sentences given on the basis of the above improper location conditions
Detecting means for detecting an inappropriate part in the input sentence; information on the type of the input device whether the input means for the given sentence is a kana-kanji conversion input device or a character recognition device
Or identification information that identifies the person who entered the information,
Input identification means for identifying at least one piece of information; and condition changing means for changing the content stored in the condition storage means in response to the notification from the input identification means. Appropriate point detection device. 2. An apparatus for detecting an incorrect part of a sentence which automatically detects and corrects an inappropriate part of a given sentence, comprising: a condition storage means for storing a condition of an inappropriate part and a correction method; Stored in the condition storage means in the sentence
The sentences given on the basis of the above improper location conditions
Detecting means for detecting an improper part in the correction means, applying a correction method to the part detected by the detecting means to generate a correction candidate, and input means for the given sentence being a kana-kanji conversion input device. Information on the type of input device, whether it is a character recognition device
Or identification information that identifies the person who entered the information,
Input identification means for identifying at least one piece of information; and condition changing means for changing the content stored in the condition storage means in response to the notification from the input identification means. Appropriate point detection device.