JPH04152458A - Word processor - Google Patents

Abstract

PURPOSE:To easily confirm the contents of a document by analyzing each word, each paragraph, and an entire sentence for each sentence, segmenting plural words showing the gists, and generating the auxiliary verbs, etc., for production of the sentences. CONSTITUTION:A syntax analyzing part 14 absorbs the auxiliary verbs, etc., into the precedent words and arranges them in the paragraphs in a word list analyzed by a morpheme analyzing part 13. Thus, the part 14 produces a reference notion structure. A semantic analyzing part 15 produces the Japanese word notion reliant structure data where the active-passive relation is adjusted against the reference notion structure. A gist segmenting part 16 segments the necessary words against the Japanese word notion reliant structure data based on a rule stored in a rule memory 19. A generating part 17 produces an auxiliary verb absorbed by the part 14 and produces a sentence. This sentence is outputted to a control part 10 and displayed at an output part 12. In such a constitution, the gist of a produced document is surely extracted and the contents of the document can be easily confirmed from the extracted gist.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a document creation device having a function of creating main points of a document.

(Prior art) In recent years, in document creation devices (word processors),
There are some that have a text editing function. When checking a created sentence using this sentence refinement function, after cutting out one sentence from the sentence, morphological analysis is performed by searching an analysis dictionary to analyze it into a series of sentences, and inappropriate parts of the sentence are removed according to predetermined rules. A replacement candidate is created, shown to the user using properties, etc., and proofreaded according to the user's instructions. When creating replacement candidates for the inappropriate portions of a sentence, conventionally the sentences are semantically analyzed and the sentences are rearranged based on the analysis results to create easy-to-understand sentences.

(Problems to be Solved by the Invention) As mentioned above, in conventional document creation devices, when presenting replacement candidates in text editing, they simply rearrange the sentences to create an easy-to-understand sentence, and only focus on the main points. I couldn't cut it out.

The present invention has been made in view of the above-mentioned points, and it is an object of the present invention to provide a document creation device that can extract only the main points of a sentence and that makes it extremely easy to select replacement candidates when revising a sentence.

[Structure of the Invention] (Means and Effects for Solving the Problems) That is, the document creation device according to the present invention includes an analysis dictionary that stores knowledge information used for Japanese language analysis, and a document creation device that refers to this analysis dictionary. A morphological analysis means that performs word-by-word analysis based on the analysis result of this morphological analysis means, a syntactic analysis means that performs clause-by-clause analysis based on the analysis result of this morphological analysis means and absorbs auxiliary verbs etc. into the previous word, a semantic analysis means for analyzing a whole sentence based on the analysis results; a gist extraction means for cutting out necessary words from the analysis result of the semantic analysis means; and a syntactic analysis means for absorbing the words extracted by the gist cutting means. and sentence generation means for generating one sentence by generating the auxiliary verbs.

As a result, the created document is created for each sentence in "word units",
Analyzes are performed sequentially for "phrase units" and "- whole sentences," and words indicating key points are extracted based on the analysis results. Then, auxiliary verbs and the like are generated for the extracted words, and a sentence indicating the main point is created. Therefore, it becomes possible to easily confirm the document content from the extracted key points.

(Example) Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the basic configuration of the document creation device of this embodiment.

In Figure 1, 10 is a microprocessor (CPU)
11 is an input section connected to the control section 10, and 12 is a display section connected to the control section 10. The input section 11 accepts inputs using various keys, a mouse, etc., and inputs them to the control section IO. The display unit 12 displays input data or processed data, and is composed of a CRT or a dot matrix type liquid crystal display device, for example, a 640×40
The size is set to 0 dot.

The control unit 10 controls the entire device, and performs document creation processing, document editing processing, display control processing, etc. by starting a program according to input instructions, and also controls the morphological analysis unit 13, the syntax analysis unit 14, the semantic analysis unit 14, and the like. The analysis section 15, main point extraction section 16, and generation section 17 are controlled to execute main point extraction control.

The morphological analysis unit 13 searches the analysis dictionary 18 for the sentences sent from the control unit 10 and creates a list of the sentences analyzed word by word. The analysis dictionary 18 stores knowledge information used for this Japanese language analysis.

The syntactic analysis unit 14 absorbs auxiliary verbs and the like into previous words in the word list analyzed by the morphological analysis unit 13 and combines them into clauses to create a reference conceptual structure.

The semantic analysis unit 15 performs analysis through the - sentence, and creates Japanese concept dependent structure data in which the dependencies are adjusted for the reference concept structure created by the syntactic analysis unit 14.

The gist extraction section 16 extracts necessary words from the Japanese concept dependent structure data based on the rules stored in one rule memory and outputs them to the generation section 17 . The rule memory 19 stores rules for cutting out the main points of sentences.

The generation unit 17 generates the auxiliary verbs absorbed by the syntax analysis unit 14, creates one sentence, outputs it to the control unit 10, and displays it on the display unit 12.

Next, the gist extraction operation in the above embodiment will be explained with reference to the flowchart of FIG.

When cutting out the main points, the user specifies the text from the input section 11 and instructs to cut out the main points while the target text is displayed on the screen. When the control unit 10 is instructed to cut out the main points, it starts the processing operation shown in the flowchart of FIG. 2. That is, when the control unit 10 is instructed to cut out the main points, it outputs the target sentence to the morphological analysis unit 13. As a result, the morphological analysis unit 13
performs the morphological analysis operation shown in step AI in FIG.

Now, for example, as shown in Figure 3, Bl, I studied a lot, so I was able to pass this exam.

When a sentence is sent from the control unit 10 to the morphological analysis unit 13, the morphological analysis unit 13 refers to the analysis dictionary 18 and performs morphological analysis. Through this morphological analysis, B2 in Figure 3
As shown in the list, analyzed word by word, I studied quite a lot, so I was able to pass the exam this time. Create 7.

Next, the control unit 10 outputs the word list analyzed by the morphological analysis unit 13 to the syntax analysis unit 14. The syntactic analysis unit 14 absorbs auxiliary verbs and the like into the previous words in the word list and combines them into clauses, creating a reference conceptual structure as shown at 83 in FIG. 3 (step A2). In other words, with the object word ``pass'' at the top, ■ ``quite'' and ``study''.

Two systems are associated: "So", ■ "This time", and "Test". Moreover, "I" is associated with "study".

The reference concept structure data created by the syntax analysis unit 14 is then sent to the semantic analysis unit 15, where its meaning is analyzed (step A3), and the dependencies are adjusted as shown in B4 of FIG. A word concept dependent structure is created. That is,
In the standard concept structure, "I" which was related to "studying" is adjusted by the semantic analysis unit 15 so that it is related to "passing",
The information is sent to the main point extraction section 16.

The gist cutting unit 16 cuts out necessary words from the Japanese concept dependent structure data created by the semantic analysis unit 15 based on gist cutting rules stored in the rule memory 19 (step A4). That is, the main point cutting section 16
As shown at 85 in the figure, ``Washi'' is the arc of the subject and objective case based on the gist extraction rules from the Japanese concept-dependent structure data.

“Pass” and “Test” are cut out and output to the generation unit 17.

The generation unit 17 generates the syntax analysis unit 14 from “I”, “pass”, and “examination” extracted by the gist extraction unit 16.
The system generates the auxiliary adjustment absorbed in the step A5, and creates one sentence "I passed the exam" as shown at 86 in FIG. 3 (step A5). The sentence generated by this generation unit 17 is
It is sent to the control section 10 and displayed on the display section 12.

The main points of the created document are extracted as described above.

[Effects of the Invention] As described above, the document creation device according to the present invention can process the created document for each sentence by "word unit", "bunse unit unit", or "bunsetsu unit".
- The entire sentence is sequentially analyzed, and based on the analysis results, multiple words that indicate the main points are extracted, and auxiliary verbs are generated for these words to create a sentence. The key points can be reliably extracted, and the content of the document can be easily confirmed from the key points.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing a document creation device according to an embodiment of the present invention, FIG. 2 is a flowchart showing the operation of the embodiment, and FIG. 3 is a diagram showing a specific example of key point extraction in the embodiment. It is. DESCRIPTION OF SYMBOLS 10... Control part, 11... Input part, 12... Display part, 13... Morphological analysis part, 14. Syntax analysis part, 15.
...Semantic analysis part, 16...Gist extraction part, 17...
- Generation unit, 18... Analysis dictionary, 19... Rule memory.

Claims (1)

[Claims] In the document creation device, there is an analysis dictionary that stores knowledge information used for Japanese language analysis, a morphological analysis means that performs word-by-word analysis by referring to this analysis dictionary, and an analysis result of the morphological analysis means. There is a syntactic analysis means that analyzes each clause and absorbs auxiliary verbs into the previous word, a semantic analysis means that analyzes the whole sentence based on the analysis results of this syntactic analysis means, and a syntactic analysis means that analyzes the entire sentence based on the analysis results of this syntactic analysis means. The present invention further comprises: gist extraction means for cutting out words; and sentence generation means for generating one sentence by generating auxiliary verbs absorbed by the parsing means for the words extracted by the gist extraction means. Characteristic document creation device.