TH65852B - Systems and procedures for indistinct sentences for languages with unclear sentence boundaries. - Google Patents

Claims (6)

Disclaimer (all) which will not appear on the advertisement page: EDIT 24/03/2017 1. Sentence cutting system and process for languages with indistinct sentence boundaries, consisting of external devices (11) that perform input functions. The continuous text enters the sentence cutter (10), the sentence cutter (10) that controls the function of the sentence cutter. Contains - An input unit (12) that receives data, contiguous text or strings from a database (17) or from an external device (11), such - a memory (16) that stores the resulting data. Received from the input unit (12) - the processor (14) that controls the operation of the system and processes the statistical value of the data that Received from the input unit (12) or from the memory (16) - the part for the output (13) that transmits the received data from the processor (14) to the display unit (19) and / or. It is transmitted to the database (17) and is unique in that the processor (14) checks the gap between each token as a cut-off position. Sentence or not The results of the statistical probability are the trigram probability of the word type and the word probability value when determining the word type. With the method of bringing the token to the previous token Page until you find a gap in the intersection position. Then send the clipped sentence to the output section (13) to display the clipping information. The processor (14) has a synthetic system. Speech to convert processed text into speech 2. Sentence cutting systems and processes for languages with unclear sentence boundaries. According to claim 1, where the method in To cut the said sentence It consists of the following steps: a) Getting a string input consisting of a certain number of characters (21) from a database (17) or from an external device (11) is unique. Complete with step b) token (22) string break, using space as a separator between the words. Space into tokens c) Building a series of words (23) by breaking the string into a series of words. By word series Has a structure consisting of Previous address token (PREVIOUS_TOKEN), space, current token (CURRENT_TOKEN), respectively, to process the word type identification by statistical method (24) d) statistical POS tagging (24) to identify the word type of the gap. By using the statistical values of words and spaces in a word series, namely the trigram probability of the word type and the word probability when determining the word type, e) the channel word type detection. Space in the word series (25) When the word type of the space is found as a sentence break, the sentence will be omitted at the gap position. And send all of the token chains preceding the gap Display or save in database (28) as trimmed sentence and search for the next space by repeating steps b) to e) until the incoming string is exhausted. 3. Cutting system and process. Break sentences for languages where sentence boundaries are not clearly visible. According to claim 1 or 2, where the sentence breakdown system for languages with indistinct sentence boundaries is Thai language. 4. Sentence cutting systems and procedures for languages with unclear sentence boundaries. According to claim 1, where external devices (11) are keyboards and / or databases, 5. Sentence cutting systems and procedures for languages with unclear sentence boundaries. According to claim 1, where the sentence-cutting device (10) is a computer microprocessor or digital signal processor. --------------- Edited 1/13/2015 1. A method for sentence breaks for languages that have inconsistent sentences and use space as the boundaries. Of each sentence, for use in a sentence cutting system that includes the memory processor, the part for the input. And a section for output Where the above method Contains the steps of a) taking a string input consisting of a number of characters from the section for such input; b) splitting the string into a number of tokens; c) Analysis of the space within each such token. It takes into account the contextual word of the gap, statistical POS tagging, to determine whether the space is the intersection position or not; With unique characteristics The process also consists of d) examining the gaps between each token to determine whether the gaps found were Sentence intersection position or not using statistical probability results; e) Continuing the token with the previous one. Also repeat steps b), c) and d) if the gap found is not the intersection position. It will repeat until Will find a gap at the intersection position; f) Output all the token as a clause clause 2. How to cut clause according to claim 1, where the language with such indistinct syntax is Thai 3. System Cut text It includes a ton of string data input units such as keyboards, databases, processors, memory, and display units. The process of cutting sentences according to claim 1 or 2. 4. System for speech synthesis of Braille text for the conversion of input into the text. The system becomes a voice for blind communication using a clause-cutting method according to claim 1 or 2. 5. Text display system for languages with indistinct sentence boundaries and using spaces as boundaries. Of each sentence Where the display system can cut the input text into The system can be sentence. By using the method according to claim 1 or 2 ---------------------------------------- ---------------- EDIT 6/2/2015 1. Sentence cutting systems and procedures for languages with indistinct sentences, consisting of external devices (11) that act as text input. Continuously into the sentence cutting device (10), the sentence cutting device (10) that controls the function of the sentence cutting system. Contains - An input unit (12) that receives data, contiguous text or strings from a database (17) or from an external device (11), such - a memory (16) that stores the resulting data. Received from the input unit (12) - the processor (14) that controls the operation of the system and processes the information received from the input unit (12) or from the memory (16), and - the part for the output (13). ) That transmits the data received from the processor (14) to the display unit (19) and / or transmits it to a database (17), characterized by the processor (14). ) Check the gap between each token as a position. Cut the sentence or not Using statistical probability results With the method of bringing that token to the previous token Page until you find a gap in the intersection position. The clipped sentence is then sent to the output section (13) to display the clipping information. 2. Sentence cutting systems and processes for languages with indistinct sentences. According to claim 1, where the processor has a system for the speech synthesis of Braille text for the conversion of the text entered into Such a system becomes the voice for communication of the blind by means of sentence cutting. 3. Sentence cutting systems and methods for languages with indistinct sentences. According to claim 1 or 2, where the method of dividing the said sentence It consists of the following steps: a) Getting a string containing a certain number of characters (21) from a database (17) or from an external device (11) b) splitting that string into a number of tokens. One (22) c) Analysis of the space within each such token. The contextual word of the gap (23) is considered by statistical POS tagging (24) to determine whether the gap is the intersection position, or D) checking the gaps between each such token so that the gaps found are The intersection position or not using the statistical probability result (25) e) the next token is applied to the previous one (26), and repeats steps b), c) and d. ) If it is found that the said gap is not the intersection position. Repeat until you find a space that is Sentence Cut Position (27) f) All the aforementioned token outputs as a clause clause (28) 4. Sentence cutting systems and procedures for languages with indistinct sentences. According to one of the claims 1 to 3, where the processor cuts the language with unclear sentence boundaries, the said language is Thai. 5. The system and process of sentence cutting for languages with unclear sentence boundaries. According to claim 1, where external devices (11) are keyboards and / or databases. 6. Sentence cutting systems and procedures for languages with unclear sentences. According to claim 1, where the sentence-cutting device (10) is a computer microprocessor or digital signal processor. ------------------------------------------- 1. Syntax cutting systems and processes for languages with indistinct sentences consist of an external device (11) that performs a continuous input of text into a sentence-cutting device (10), a sentence-cutting device (10). Function to control the function of the sentence cutting system. Contains - An input unit (12) that receives data, contiguous text or strings from a database (17) or from an external device (11), such - a memory (16) that stores the resulting data. Received from the input unit (12) - the processor (14) that controls the operation of the system and processes the information received from the input unit (12) or from the memory (16), and - the part for the output (13). ) That transmits the data received from the processor (14) to the display unit (19) and / or transmits such information to the database (17), whose characteristics are the approximation (14). Check the gap between each token as a position. Cut the sentence or not Using statistical probability results With the method of bringing the token to the previous token Page until you find a gap in the intersection position. Then send the clipped sentence to the output section (13) to display the clipping information. 2. Sentence cutting systems and processes for languages with unclear sentence boundaries. According to claim 1, where the processor has a system for speech synthesis from Braille text for the conversion of the text entered into The system becomes a speech for the blind by means of sentence cutting. 3. Sentence cutting systems and procedures for languages with unclear sentence boundaries. According to claim 1 or 2, where the method of dividing the said sentence The following steps are included: a) Getting string input consisting of a certain number of characters (21) from a database (17) or from external devices (11) b) splitting that string into tokens. A) number (22) c) Analysis of the space within each such token. The contextul word of the gap (23) is considered by statistical POS tagging (24) to determine if the gap is the intersection position, or D) checking the gaps between each such token to determine whether the gaps found were The intersection position or not, using statistical probability results (25) e), the cocaine braning continues with the preceding cocaine (26) and repeats steps b), c) and D) If it is found that the said gap is not the intersection position. It will repeat until you find an empty space. The clause (27) f) outputs all of the token as a clipped sentence (28). 4. Sentence cutting systems and processes for languages with unclear sentence boundaries. According to claims 1 or 3, where the processor cuts the language with indistinct sentence boundaries, such is Thai. 5. Sentence cutting systems and procedures for languages with unclear sentence boundaries. According to claim 1, where external devices (11) are keyboards and / or databases. 6. Sentence cutting systems and processes for languages with unclear sentence boundaries. According to claim 1, where the sentence-cutting device (10) is a computer microprocessor or digital signal processor.