CN102867049B - Chinese PINYIN quick word segmentation method based on word search tree - Google Patents
Chinese PINYIN quick word segmentation method based on word search tree Download PDFInfo
- Publication number
- CN102867049B CN102867049B CN201210332072.7A CN201210332072A CN102867049B CN 102867049 B CN102867049 B CN 102867049B CN 201210332072 A CN201210332072 A CN 201210332072A CN 102867049 B CN102867049 B CN 102867049B
- Authority
- CN
- China
- Prior art keywords
- node
- word
- search
- tree
- search tree
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention discloses a Chinese PINYIN quick word segmentation method based on a word search tree. The method is implemented by a computer or embedded mobile equipment and comprises the following working steps of: 1, building a Chinese character PINYIN search tree according to all the known Chinese character PINYIN lists; 2, combining the search tree with a hash table according to the built word search tree, and segmenting a string of given Chinese PINYINs; 3, working out a word segmentation result; and 4, destroying the search tree and releasing resources. Due to a public prefix of a character string, a construction space is saved, so that unnecessary character string comparison is greatly reduced; by the redundancy hash table with an index, the search efficiency is improved; and the time complexity of an algorithm is reduced to the minimum.
Description
Technical field
The invention belongs to computing machine or various hand-held embedded movable equipment Chinese information processing technical field, particularly a kind of Chinese phonetic alphabet fast word segmentation method realized based on word lookup tree.
Background technology
From a string continuous print Chinese phonetic alphabet, automatically identifying each individual character phonetic by computer software algorithm, is the technology that spelling input method and search engine (associating Chinese sentence again according to phonetic type key word) must use.Using existing all Chinese word phonetics as key word, set up a Hash table, by repeatedly searching from the Hash table set up and mating during participle, a string continuous print Chinese phonetic alphabet can be realized to carry out participle, but the method exists the not high problem of efficiency.
For raising the efficiency, in prior art, following improvement is made to above-mentioned Hash table: using the initial of Chinese word phonetic as key word, set up a Hash table, each element of Hash table is a single-track link table, and store with Hash table key word letter in chained list is all individual character phonetics started.Like this after improvement, when searching at every turn, from Hash table, first obtain the first node pointer of a single-track link table according to initial fast, and then travel through single-track link table, make final coupling.The Hash table through improvement is used to improve participle efficiency, but at process polysemant, such as during " xian " " piao ", still can there is the problem needing special processing, after such as retrieving " xi ", a kind of scheme " xi " is removed from word string immediately, next continue to retrieve " an ", but phonetic " xian " just lost like this; First scheme retains " xi " in word string, continue to retrieve the phonetic started with alphabetical x, until find the end of single-track link table, and then remove " xi ", last continuation again retrieves " an ", so just can find all possible phonetic plan " xi " " an " " xian ", but this scheme search efficiency is lower.
Chinese patent (the patent No.: 200710118921), a kind of internal memory processing method of telephone number mapping domain name server and device, although this patent refer to search tree and Hash table, but this patent is only the use of the basic function that search tree and Hash table store and search, not through any improvement, also without any extension and improvement; And also fundamental difference is there is in the purposes used, this patent just stores data on the node of search tree and Hash table, the just simple function being found the data of storage by search tree and Hash table, and the present invention mainly realizes is complete Chinese phonetic alphabet fast word segmentation by the combination of a kind of mutation of search tree and Hash table, this patent is the data of searching storage, the present invention has been the fast word segmentation of the Chinese phonetic alphabet, and two files exist the difference of essence in use-pattern and purposes.
Chinese patent (the patent No.: 200810129141.8), the method and apparatus of adjustment candidate word sequence, although this patent refer to search tree and Hash table, but the trie tree in this patent has just used the one in search tree or Hash tree, as in the present invention search tree and Hash table are not combined use, both combine closely in the present invention, indispensable; And purposes is also different, the method and apparatus of this patent adjustment candidate word sequence just utilizes search tree or Hash table to store, judge whether the pinyin string stored is standard spelling, search tree or Hash tree is used to be not the function being used as participle, and the present invention uses the combination of the mutation of search tree and Hash tree to complete the fast word segmentation of phonetic, finally form pinyin string.
Chinese patent (the patent No.: 200910107961.1), a kind of implementation method of fast word segmentation, although this patent is also a kind of method of participle, but this searching patents tree is realized by one-level concordance list and HASH multiway tree, the deficiency of this patent is: when processing polysemant, the situation of easy Zao Cheng Lost word; If do not want the situation occurring Lost word, just need the method taking to reduce search efficiency to solve, although ensure that the correctness of lookup result, most direct result is exactly cause the lower problem of search efficiency.
Although the single-track link table in such scheme has easy care in such as internal memory, inserts and delete the advantage such as simple, have that query performance is low, the shortcoming of search efficiency Di Huo Lost word.
Summary of the invention
Object of the present invention is exactly to solve the problem, a kind of Chinese phonetic alphabet fast word segmentation method realized based on word lookup tree is provided, search tree is combined with Hash table, a kind of mutation of use Hash tree has carried out the fast word segmentation to the Chinese phonetic alphabet, this kind of participle mode had both avoided that query performance is low, efficiency Di, Lost word problem, turn improve search efficiency, achieve fast word segmentation.
To achieve these goals, the present invention adopts following technical scheme:
Based on the Chinese phonetic alphabet fast word segmentation method that word lookup tree realizes, the method is realized by computing machine or embedded movable equipment, and groundwork step is as follows:
Step one, set up Chinese word phonetic search tree according to all known Chinese word pinyin tables;
The word lookup tree that step 2, foundation have been set up, is combined search tree with Hash table, carries out participle to given a string Chinese phonetic alphabet;
Step 3, provide word segmentation result;
Step 4, destruction search tree, releasing resource.
In described step one, set up individual character phonetic search tree according to all known Chinese word pinyin tables, mainly comprise following steps:
(1) root node does not comprise character, and except root node, each node only comprises a character;
(2) character that comprises of all child nodes of each node is not identical;
(3) except leaf node, each node has a length to be the Hash table of 26, and Hash table is with the ascending order of 26 English alphabets for index, and each element stores a child node respectively, and child node actual number is less than or equal to 26;
(4) each node comprises an identification field, this field value 0 or 1, and be used for mark from root node to this node, on path, the Connection operator of process gets up whether to represent a complete Chinese word phonetic.
In described step 2, according to the word lookup tree of having set up, search tree is combined with Hash table, participle is carried out to given a string Chinese phonetic alphabet, mainly comprises following steps:
A) once search for from root node;
B) obtain the first letter will searching keyword, and from Hash table, select corresponding subtree according to this letter and forward this subtree to proceed retrieval;
C) in corresponding subtree, obtain the second letter will searching keyword, and select corresponding subtree to retrieve further;
D) iterative process: obtain the 1st, 2 of keyword the ... n letter, continues to search.
In described step 3, concrete steps are as follows:
1) at certain node place, if all letters of keyword have been removed or the identification field value of node is 1, then from root path to present node, the identification field values of all characters and present node has been exported successively.
2) if all letters of keyword are removed, namely complete and search; Otherwise obtain the next one letter of keyword, the root node getting back to search tree continues iterative search.
3) for the keyword comprising multiple semanteme, such as: piao, both can be construed to " ticket ", and also can be construed to " fur-lined jacket ", word segmentation result will export all possible value.
In described step 4, after completing participle, destroy search tree, releasing resource, reclaim the internal memory taken.
Described search tree is the polymorphic set of a kind of support, the data structure that comprise insertion, delete and search etc. operates.
Described Hash table, is also named hash table, is the data structure directly conducted interviews according to key value.
Beneficial effect of the present invention:
This invention is a new breakthrough at computing machine or various hand-held embedded movable equipment Chinese information processing technical field, and comprehensive Xie Jue Lost word, the problem that search efficiency is low, provide a kind of new approaches of Chinese information processing technical field.
The present invention utilizes the common prefix of character string to save structure space, reduce meaningless character string comparison to greatest extent, not only increase the search efficiency of phonetic like this, also effectively can improve the internal memory service efficiency of computing machine or various hand-held embedded movable equipment, effectively improve the operational efficiency of various equipment;
The present invention utilizes the redundancy Hash table of tape index to improve search efficiency, reduces the time complexity of algorithm to greatest extent.Effectively reduce the number of searches of node, effectively improve the real-time of algorithm, and ensure that the accuracy of searching.Meanwhile, also effectively improve search efficiency and effectively save query time.
Participle mode of the present invention had both avoided that query performance is low, efficiency Di, Lost word problem, turn improved search efficiency, achieved fast word segmentation.Participle technique is a new breakthrough, because search efficiency improves, participle efficiency improves, and also can improve the work efficiency of Chinese phonetic alphabet user accordingly, saves working time, reduces labour intensity.
Accompanying drawing explanation
Fig. 1: search tree participle process flow diagram;
Fig. 2: the structural drawing of search tree;
Fig. 3: participle process flow diagram flow chart.
Embodiment
Below in conjunction with accompanying drawing and embodiment, the invention will be further described.
As shown in Figure 1, first set up the Hash tree of a search tree and Hash table combination according to existing Chinese word pinyin table, then participle is carried out to given a string continuous print Chinese phonetic alphabet, provide analysis result, finally destroy search tree, releasing resource, reclaim internal memory.
Set up Hash tree, set up a word lookup tree according to all known Chinese word pinyin tables.The root node of search tree does not comprise character, and except root node, each node only comprises a character.The character that all child nodes of each node of search tree comprise is not identical.Search tree is except leaf node, and each node has a length to be the Hash table of 26, and Hash table is with the ascending order of 26 English alphabets for index, and each element stores a child node respectively, and child node actual number is less than or equal to 26.Search tree is except root node, and each node comprises an identification field, this field value 0 or 1, and be used for mark from root node to this node, on path, the Connection operator of process gets up whether to represent a complete Chinese word phonetic.
Word segmentation processing, according to the word lookup tree of having set up, carries out participle to given a string Chinese phonetic alphabet.Once search for from root node.Obtain the first letter will searching keyword, and from Hash table, select corresponding subtree according to this letter and forward this subtree to proceed retrieval.In corresponding subtree, obtain the second letter will searching keyword, and select corresponding subtree to retrieve further.Same step, obtains the 3rd of keyword, the 4th, the n-th letter, and retrieves corresponding subtree.
Export word segmentation result, at certain node place, all letters of keyword have been removed or the identification field value of node is 1, then from root path, export the identification field values of all characters and present node to present node successively.If all letters of keyword are removed, namely complete and search; Otherwise obtain the next one letter of keyword, the root node getting back to search tree continues iterative search.For the keyword comprising multiple semanteme, such as: piao, both can be construed to " ticket ", and also can be construed to " fur-lined jacket ", word segmentation result will export all possible value.
Destroy Hash tree, after completing participle, destroy Hash tree, reclaim the internal memory taken.
As shown in Figure 2, as space is limited, the phonetic transcriptions of Chinese characters of part of representative is only listed in figure.
" a ", root node comprises an identification field (value 0) and a Hash table, from first element (representing the child node that character is a) of root node Hash table down, (character is a) comprise an identification field too to first order child node, at this moment identification field value is 1, because namely alphabetical a represents a complete Chinese phonetic alphabet one.
" ai " " an " " ao ", continuation character is the node of a, it also has a Hash table, because the not only " " phonetic started with alphabetical a equally, also have " ai " " an " " ao ", so a node has character to be three child nodes of " i " " n " " o ", and the attribute field of three child nodes all value be 1.
" ang ", equally, character is that in Hash table, also contain a character be the child node of g to the node of n, and the zone bit of child node is also 1.So by the path from root node through a – n – g connect after obtain phonetic " ang ".
" pi " " po " " pian ", derivation principle is the same.
As shown in Figure 3, in this process flow diagram, part uses the false code of class C to carry out descriptor flow process.
Whole process flow diagram, describes complete participle process, is subdivided into a main flow and three sub-processes.Comprise the polysemant method in process pinyin string, during process polysemant, employ recursive call mode.
Starting of flow process is input a string Chinese phonetic alphabet and initialization local variable, and pc1 is used for the beginning of a record individual character participle; Pc2 is a dynamic cursor, mark current character; Pc3 is used for the end of a record individual character participle, and object is to process polysemant; Pt is a dynamic cursor, mark current lookup tree node.
Main flow is an iterative process, and the value according to the child node of the zone bit of pt node, character that pc2 points to and pt node carries out a series of judgement, and then jumps to different sub-processes.
The job step of main flow is:
Step 1, participle pinyin string is treated in input;
Step 2, statement following variables: (1) character pointer pc1, pc2, pc3, (2) search tree node pointer pt; Pc1=pc2 points to the first character pc3=null of pinyin string;
Step 3, the root node of pt=search tree;
Step 4, judges whether the identification field of pt node equals 1, if just enter sub-process 1, just enters step 5 if not;
Step 5, judges whether the character that pc2 points to is empty, if just enter sub-process 2, just enters step 6 if not;
Step 6, judge with pc point to character as index, can child node be found in the Hash table of pt node; If just enter step 7; Otherwise enter sub-process 3;
Step 7, performs pc2++, pt=child node; Return step 4;
Sub-process 1, when the flag of pt node is 1, representative have found a complete individual character phonetic, now exports this individual character phonetic, then according to pc3(polysemant token variable) value continue Recursion process polysemant, main flow is returned in last redirect.
The detailed step of sub-process 1 is:
Step (1-1), exports all characters pointed to from pc1 to pc2-1, and reports that this is an individual character phonetic;
Step (1-2), judges whether pc3 is empty, if just enter step (1-4), just enters step (1-3) if not;
Step (1-3), intercept all characters pointed to from pc3 to pc2-1, recursive call from this flow process, enters step (1-4);
Step (1-4), pc3=pc2; Return the step 5 of main flow.
When the character that sub-process 2, pc2 is pointed to is empty, representative has arrived the ending treating participle pinyin string, if now pc1(represents the beginning of an individual character participle) be not equal to pc2, export all characters that pc1 to pc2-1 points to, and be reported as unallowable instruction digit string, finally jump out whole participle flow process.
The detailed step of sub-process 2 is:
Step (2-1) judges whether pc1 equals pc2, if just terminate, just enters step (2-2) if not;
Step (2-2) exports all characters pointed to from pc1 to pc2-1, and reports that this is unallowable instruction digit string, then terminates.
Sub-process 3, at pt(current lookup tree node) child node in do not find the node of coupling, now according to pc3(polysemant token variable) value do further judgement, if pc3 is null value, then representative encounters illegal character (such as character i, it is not the beginning character of any Chinese word phonetic), exports this unallowable instruction digit, after resetting variable, jump to the beginning of main flow; Otherwise representative needs the position (, in the beginning of string, pc2 and pc3 is in the position of second p for such as pipi, present pc1) getting back to pc3, is, after resetting variate-value, jump to the beginning of main flow equally.
The detailed step of sub-process 3 is:
Step (3-1): judge whether pc3 is empty, if so, enters step (3-2); If not, step (3-6) is just entered;
Step (3-2): export all characters pointed to from pc1 to pc2, and report that this is unallowable instruction digit string; Enter step (3-3);
Step (3-3): judge whether pc1==pc2 sets up, if so, just enters step (3-4), if not, just enters step (3-5);
Step (3-4): pc2++; Enter step (3-5);
Step (3-5): pc1=pc2; Enter the step 3 of main flow;
Step (3-6): pc1=pc3; Pc2=pc3; Pc3=null; Enter the step 3 of main flow.
By reference to the accompanying drawings the specific embodiment of the present invention is described although above-mentioned; but not limiting the scope of the invention; one of ordinary skill in the art should be understood that; on the basis of technical scheme of the present invention, those skilled in the art do not need to pay various amendment or distortion that creative work can make still within protection scope of the present invention.
Claims (5)
1., based on the Chinese phonetic alphabet fast word segmentation method that word lookup tree realizes, the method is realized by computing machine or embedded movable equipment, it is characterized in that, groundwork step is as follows:
Step one, set up Chinese word phonetic search tree according to all known Chinese word pinyin tables, mainly comprise following steps:
(1) root node does not comprise character, and except root node, each node only comprises a character;
(2) character that comprises of all child nodes of each node is not identical;
(3) except leaf node, each node has a length to be the Hash table of 26, and Hash table is with the ascending order of 26 English alphabets for index, and each element stores a child node respectively, and child node actual number is less than or equal to 26;
(4) each node comprises an identification field, this field value 0 or 1, and be used for mark from root node to this node, on path, the Connection operator of process gets up whether to represent a complete Chinese word phonetic;
The word lookup tree that step 2, foundation have been set up, is combined search tree with Hash table, carries out participle to given a string Chinese phonetic alphabet;
Step 3, provide word segmentation result;
Step 4, destruction search tree, releasing resource.
2. as claimed in claim 1 a kind of based on word lookup tree realize Chinese phonetic alphabet fast word segmentation method, it is characterized in that, in described step 2, according to the word lookup tree of having set up, search tree is combined with Hash table, participle is carried out to given a string Chinese phonetic alphabet, mainly comprises following steps:
A) once search for from root node;
B) obtain the first letter will searching keyword, and from Hash table, select corresponding subtree according to this letter and forward this subtree to proceed retrieval;
C) in corresponding subtree, obtain the second letter will searching keyword, and select corresponding subtree to retrieve further;
D) iterative process: obtain the 1st, 2 of keyword the ... n letter, continues to search.
3. as claimed in claim 1 a kind of based on word lookup tree realize Chinese phonetic alphabet fast word segmentation method, it is characterized in that, in described step 3, concrete steps are as follows:
1) at certain node place, if all letters of keyword have been removed or the identification field value of node is 1, then from root path to present node, the identification field values of all characters and present node has been exported successively;
2) if all letters of keyword are removed, namely complete and search; Otherwise obtain the next one letter of keyword, the root node getting back to search tree continues iterative search;
3) for the keyword comprising multiple semanteme, word segmentation result is exported.
4. a kind of Chinese phonetic alphabet fast word segmentation method realized based on word lookup tree as claimed in claim 1, is characterized in that, in described step 4, after completing participle, destroy search tree, releasing resource, reclaim the internal memory that takies.
5. a kind of Chinese phonetic alphabet fast word segmentation method realized based on word lookup tree as claimed in claim 1, is characterized in that, described search tree a kind of supports polymorphic set, comprises the data structure of insertions, deletion and search operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210332072.7A CN102867049B (en) | 2012-09-10 | 2012-09-10 | Chinese PINYIN quick word segmentation method based on word search tree |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210332072.7A CN102867049B (en) | 2012-09-10 | 2012-09-10 | Chinese PINYIN quick word segmentation method based on word search tree |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102867049A CN102867049A (en) | 2013-01-09 |
| CN102867049B true CN102867049B (en) | 2015-02-25 |
Family
ID=47445918
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210332072.7A Active CN102867049B (en) | 2012-09-10 | 2012-09-10 | Chinese PINYIN quick word segmentation method based on word search tree |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102867049B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105955986A (en) * | 2016-04-18 | 2016-09-21 | 乐视控股（北京）有限公司 | Character converting method and apparatus |
| CN110019649A (en) * | 2017-12-25 | 2019-07-16 | 北京新媒传信科技有限公司 | A kind of method and device established, search for index tree |
| CN109947779B (en) * | 2019-03-29 | 2021-01-22 | 北京金山安全软件有限公司 | Storage method, device and equipment for user input vocabulary |
| CN111310450B (en) * | 2020-03-23 | 2023-07-14 | 中国建设银行股份有限公司 | Character string word segmentation method, device, equipment and storage medium |
| CN111444413B (en) * | 2020-04-08 | 2023-05-12 | 作业不凡(北京)教育科技有限公司 | Data query method and device and computing equipment |
| CN113065340A (en) * | 2021-03-11 | 2021-07-02 | 国网河北省电力有限公司 | Similar asset prompting method and device based on similar word mining |
| CN113094470B (en) * | 2021-04-08 | 2022-05-24 | 蔡堃 | Text searching method and system |
| CN113961768B (en) * | 2021-12-22 | 2022-05-13 | 金蝶软件（中国）有限公司 | Sensitive word detection method and device, computer equipment and storage medium |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6648120B2 (en) * | 1998-05-19 | 2003-11-18 | Ssd Control Technology, Inc. | Device for clamping and holding an elongated workpiece |
| CN101751430A (en) * | 2008-12-12 | 2010-06-23 | 汉王科技股份有限公司 | Electronic dictionary fuzzy searching method |
| CN101833547B (en) * | 2009-03-09 | 2015-08-05 | 三星电子（中国）研发中心 | The method of phrase level prediction input is carried out based on individual corpus |
| CN101576877A (en) * | 2009-06-16 | 2009-11-11 | 程治永 | Fast word segmentation realization method |
| CN101699438B (en) * | 2009-11-04 | 2013-04-17 | 北京锋力信息科技有限公司 | Data access method and system |
-
2012
- 2012-09-10 CN CN201210332072.7A patent/CN102867049B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN102867049A (en) | 2013-01-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102867049B (en) | Chinese PINYIN quick word segmentation method based on word search tree | |
| US10169426B2 (en) | Fast identification of complex strings in a data stream | |
| US8301437B2 (en) | Tokenization platform | |
| Baeza-Yates | Introduction to Data Structures and Algorithms Related to Information Retrieval. | |
| US20090063538A1 (en) | Method for normalizing dynamic urls of web pages through hierarchical organization of urls from a web site | |
| CN103365992B (en) | Method for realizing dictionary search of Trie tree based on one-dimensional linear space | |
| CN107153647B (en) | Method, apparatus, system and computer program product for data compression | |
| JP2000231563A (en) | Document retrieving method and its system and computer readable recording medium for recording document retrieval program | |
| CN104063384A (en) | Data retrieval method and device | |
| CN105589894B (en) | Document index establishing method and device and document retrieval method and device | |
| CN110795526A (en) | Mathematical formula index creating method and system for retrieval system | |
| CN114443646A (en) | Bi-directional linear linked list directory retrieval based multi-way tree structure, electronic device and readable storage medium | |
| CN111190873B (en) | Log mode extraction method and system for log training of cloud native system | |
| CN112069305B (en) | Data screening method and device and electronic equipment | |
| CN106250456A (en) | Bid winning announcement extraction method and device | |
| CN108549679B (en) | File extension fast matching method and device for URL analysis system | |
| CN103116607B (en) | A kind of text retrieval system based on the Chinese phonetic alphabet newly | |
| KR101089722B1 (en) | Method and apparatus for prefix tree based indexing, and recording medium thereof | |
| CN105426490A (en) | Tree structure based indexing method | |
| Kanda et al. | Practical rearrangement methods for dynamic double‐array dictionaries | |
| CN109657108B (en) | Domain name asset data storage and query method and system | |
| CN108595584B (en) | Chinese character output method and system based on digital marks | |
| CN105095276B (en) | Method and device for mining maximum repetitive sequence | |
| CN112883704A (en) | Big data similar text duplicate removal preprocessing method and device and terminal equipment | |
| CN102722527B (en) | Full-text search method supporting search request containing missing symbols |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 250101 A203, block F, entrepreneurship square, Qilu Software Park, No.1 Shunhua Road, high tech Development Zone, Jinan City, Shandong Province Patentee after: Conway Communication Technology Co., Ltd Address before: 250101 Qilu Software Park, No. 1 Shunhua Road, Jinan High-tech Development Zone, Shandong Province, A 203 Patentee before: SHANDONG KANGWEI COMMUNICATION TECHNOLOGY Co.,Ltd. |