? With the wide application of computers in the field of news and book publishing industry, various electronic publications such as the spring class continue to emerge, especially with the development of Internet technology, online e-magazines, business web pages, social networking platforms are more numerous. Whether it is a publisher's newspaper or book, or a variety of electronic information on the Internet, generally through the keyboard entry, OCR recognition or voice recognition and other ways to enter the computer. These entry methods in the current state of the art can not guarantee the accuracy of the information entered. Text proofreading has become an important part of the audit before the publication of newspapers, magazines and books, which directly affects the quality of publications. With the rapid development of the publishing industry in recent years the volume of business and electronic, proofreading link workload has increased greatly, making the traditional manual proofreading method has become more and more bottlenecks in the automation of printing and publishing, to solve the problem of the accuracy of the information entered into the emergency has become a delay.
? Therefore, the use of computers to replace the human text proofreading work is of great significance, computer proofreading than manual proofreading has obvious advantages, mainly reflected in the following aspects:
? (1) computer proofreading speed, high efficiency, not fatigue. Proofreading work itself is a more mechanical work, proofreading staff for a long time on the ground to the densely packed Chinese characters, letters, punctuation and a variety of equations, eye and spirit are very fatigue, often in the spirit of an annoyed mood, the breadth and stability of attention are plummeting, and if there is no good spirit of dedication, or even hastily measured and over, the error is also unknowingly hidden down to the books, the quality of proofreading. The quality of proofreading to the book has an impact. Computer proofreading does not exist fatigue and irritability, and its speed and efficiency is more than manual proofreading.
? (2) the computer does not have a work attitude, psychological emotions and other issues, manual proofreading due to the different personnel working environment, salary and other differences will affect the proofreader's work attitude or psychological emotions, resulting in the impact of the quality of text proofreading, and this due to the work attitude or psychological emotional problems, in the computer does not exist.
? (3) computer proofreading software in the thesaurus and terminology thesaurus capacity is very large, non-general manual proofreader's knowledge can be compared to, and proofreading of different professional manuscripts, can be connected to a different terminology thesaurus, so the Chinese text, words, grammatical errors, does not conform to the Chinese syntax and semantics of the word with the error, the leader of the person's name and the position with the error, the use of units of scientific measurements, the use of irregular, paired Errors in the use of scientific units of measurement, the wrong use of punctuation, certain numerical errors, terms that do not conform to the professional thesaurus and spelling errors in English words can be quickly found and marked in red. In addition, those errors that are easily overlooked by manual proofreading, such as "sprint" and "sprint" (error), "竟争" and "意争" (error), "shock" and "shock" (error), can be quickly identified and red-flagged. "Shock" and "shock" (wrong), "airiness" and "emotion" (wrong), "governance" and "governance" (wrong), "governance" and "governance" (wrong), "governance" and "governance" (wrong), "governance" and "governance" (wrong), "governance" and "governance" (wrong), "governance" and "governance" (wrong). "metallurgy" (error), "has" and "by" (error), etc., etc., the computer can quickly and accurately find out.
Intelligent Chinese text proofreading proofreading system composition:
? Intelligent Chinese text proofreading system mainly includes four main modules: knowledge acquisition module, preprocessing and word division module, automatic error checking module and automatic error correction module, as well as preprocessing knowledge base, error checking knowledge base, error correction knowledge base and other knowledge base systems. The relationship between each module is shown in Figure 1:
? (1) Knowledge Acquisition Module: Linguistic statistical knowledge is acquired from a large-scale corpus (including raw and cooked corpus), which is used to establish linguistic models and algorithms for text auto-checking and auto-correction. The knowledge base consists of two parts: error checking knowledge base and error correction knowledge base. The error checking knowledge base is mainly used for text error checking models and algorithms, including word frequency vector table, binary and ternary word-word congruence rate table obtained from the raw corpus, word frequency vector table, binary word congruence table, lexical binary and lexical ternary congruence table, binary meaning class and meaning class in class congruence table, and syntax knowledge base and political rule base. knowledge base and political rule base. The Error Knowledge Base is mainly used to give error suggestions for red-flagged errors, and includes confusable lexicons, similar-code word lexicons, word-driven bi-directional lexicons, English word skeleton key lexicons, and likelihood matching rules. When performing the ranking of error correction suggestions, the knowledge of word succession (obtained from the same present data) and lexical succession statistics in the error checking knowledge base is also used.
? This part is used independently of the system to obtain statistical knowledge from the corpus, and is not closely programmatically connected to the other three parts.
? (2) preprocessing and word separation module: The preprocessing and word separation module is mainly to proofread the text for word separation, at present, our system can recognize the plain text format (DXT) and rich text format (RF), for other formats of text files, such as Word, PDF, WPS and Huaguang format, you need to formatting conversion, to remove the control characters, to generate plain text format. generate a plain text format. The word is the basis of most of the self-fueled language processing system, this system is no exception, we have implemented the maximum matching of the word module, the module at the same time with the recognition of people's names, place names, due to the use of plug-in structure, this system can be completely used for the experiments of word models and algorithms, but also can be used to access the existing effect of the better word program conveniently into our system, for the checking of the model and the correction of the error model.
? (3) Automatic Error Checking Module: This module mainly implements various error checking models and algorithms, and the main function of the module is to carry out error checking for Chinese text error detection, specifically including the detection of word-level, syntactic-level, semantic-level and political errors in Chinese text, for word-level errors, we mainly use the word-level error in the Chinese text of the For word-level errors, we mainly use a combination of rules and statistics to detect errors based on the classification ideas of "non-multi-word errors" and "true multi-word errors" in Chinese texts; for syntactic-level errors, we use a combination of grammatical dictionary and statistics to detect errors based on syntactic rules and grammatical dictionaries; for semantic-level errors, we use a combination of grammatical dictionaries and statistics to detect errors; and for semantic-level errors, we use a combination of rules and statistics to detect errors. For syntactic level errors, based on syntactic rules and grammatical lexicon, a syntactic lexicon combined with statistics is used for error detection; for semantic level errors, based on the theory of denotations, a combination of semantic collocation knowledge base and the theory of evidences is used for error detection; and for political errors, based on a political rule base, a knowledge-based reasoning method is used for error detection. The output of this part is the text which is labeled for the error strings, and the result is displayed on the screen after being marked by the red sub-processes .
? (4) Automatic Error Correction Module: This module mainly implements the generation algorithm and sorting algorithm of error correction suggestions, which are based on the causes of errors in this project. For Zhuyin errors, the two-way pinyin matching method is used to locate errors and generate suggestions for error correction within a specific size of the sliding window; for Wubi errors, the similarity code computation method is used to solve the problem of locating errors and generating suggestions for error correction on the basis of a specific likelihood matching rule. For the sorting of error correction suggestions, this project constructs a sorting model for error correction suggestions based on semantic juxtaposition theory and contextual context, and determines the prioritization value of each error correction suggestion by fusing contextual information, large-scale corpus and coding information, and then uses the sorting algorithms of fast classification or bubbling to sort the correction suggestions after the prioritization value has been determined.
? Intelligent three-dimensional warehouse through the three-dimensional warehouse, automated three-dimensional warehouse two stages of development evolved, its development history is shown in the following chart:
? Intelligent three-dimensional warehouse system set of computer information management, computer control technology and mechanical engineering in one, used to solve the problem of low utilization of warehousing in the field of logistics, occupying more land, low logistics efficiency. It has a broad application prospect in the fields of mechanical parts manufacturing, medicine, tobacco, FMCG, e-commerce and so on. This project through the school-enterprise cooperation, with the support of relevant scientific research projects, for the intelligent three-dimensional warehouse system in the WMS, WCS and logistics equipment in the key technical issues carried out long-term in-depth research.
? Through continuous in-depth research, we have determined the three-layer architecture of the software part of the intelligent three-dimensional warehouse, as shown in the following figure:
?WMS management system is the core of the warehouse automation management system, which includes a series of management functions, such as warehouse information management, inventory management, inbound and outbound management and reporting, etc., and the structure of WMS management system functional modules is shown in the following figure:
? The scheduling system is responsible for issuing scheduling instructions for various hardware equipment such as stacker cranes, conveyors, forklifts, etc. The structure of the scheduling system is shown as follows: