TWI459221B - The Chinese Address Translation Method of Hierarchical Architecture - Google Patents

Description

Chinese address translation method for hierarchical architecture

The invention belongs to the fields of telecom value addition, BSS platform and technology, data mining and analysis and integrated technology.

In the past, most of the address conversions used the exhaustive method. This method must establish a dictionary to cover all administrative district names and their new administrative district names and a large number of forwarding rules. For the conversion of a single address, it is necessary to search the dictionary for its correspondence. The name of the administrative district, and according to the forwarding rules, remove the addresses that are not satisfied after the conversion one by one until you get the most qualified address.

There are two major shortcomings in the exhaustive method. The first drawback is that the execution efficiency is poor. Because of the huge dictionary, there may be thousands of combinations of the combination of numbers and national characters. Then the address is compared with many rules one by one until it is found. The most suitable re-addressing, therefore, for the transfer of large amounts of data, the operation time is very impressive; the second shortcoming is that the implementation is not effective, if the exhaustive method contains a combination is not complete, it will inevitably lead to the inability to comprehensively Corrected to the correct address, for example, there are many combinations of "Taipei County" in free format, "Taipei County", "Taipei County", "Moss County", "North County", "Taibei County", etc. Because of the diversity of manual input, a new case can be created at any time, and it is almost impossible to put a very complete case in the database.

Obviously, the above exhaustive law does not guarantee the correctness and completeness of the address after the transfer, resulting in an incorrect address after the transfer, resulting in the failure of the letter to be delivered. Loss of the rights of users and businesses.

In the patent technology 092129073 customer address multi-mode input and five-code postal area code conversion tool, there is a reference to the concept of forwarding but does not deal with free-form addresses, and is very different from the free-form address practice of this system, patent technology 092129073 is The range of the address is gradually reduced according to the previously selected area, and the zip code is converted by the code of the defined address. However, non-free-form addresses are better in integrity and correctness than free-form addresses, and the purpose of this system is to correct innately diverse addresses to more complete and valid addresses to compensate for the incorrect filling of free-form addresses. situation.

It can be seen that the above-mentioned conventional techniques still have shortcomings and need to be improved.

In view of the shortcomings derived from the above-mentioned conventional technologies, the inventor of the present invention has improved and innovated, and after years of painstaking research, he finally succeeded in researching and developing the method of converting the Chinese address into a hierarchical structure.

The object of the present invention is to provide a high-efficiency and dictionary-free addressing mechanism, so that in the process of forwarding, it is not necessary to spend a large amount of time to compare addresses one by one, and in addition to the probability of counting Chinese address architecture characteristics. Perform a quick search of address keywords so that the transfer job can be completed at high speed.

The design method for achieving the above object includes two major parts: the first part is to perform string cutting processing for each region level through the Chinese address keyword quick search program, wherein the establishment of the regional level and the keyword can be statically set. Or dynamically change; the second part is to use the conversion rule to correct the program to the first Some of the extracted strings are corrected, and the correct string values can be provided by the target database or a known database.

A Chinese-language address translation method for hierarchical architecture, the data source is a free format, and stores a string with continuous and hierarchical relationship, including the following steps: a. first setting a keyword of the current regional level according to a statistical method a probability value appearing at each position in the string; b. calculating a sum of the probability values of each of the keywords at the current level of the region at each position, if the higher position character in the sum of the probability values is included in the current region The keyword of the level, which is the boundary value of the current level of the region, otherwise it will continue to search for the location where the probability value is the second highest and is included in the current level of the region; c. Finally, according to the region level The boundary value is cut to obtain a string of each of the region levels.

The region level is a keyword with a hierarchical relationship, and the establishment of the region hierarchy and the keyword is statically set or dynamically changed, and the probability value is a target database or a matching database, and the target data is The library is a storage space for storing the data to be processed, and the known database is provided by a third party, wherein the third party provides a database of post office addresses, and the target database field includes a postal area code and an address. It is known that the fields in the database are not limited to the postal area code, the address and the area level, and further include processing the output layer level string and correcting the string to the correct string value, and the correct word The string value is provided by the target database or the known database.

The method provided by the present invention has the following advantages when compared with other conventional technologies:

1. The invention is characterized by avoiding storing a large number of dictionaries on the local end and solving the time-consuming transfer problems of the conventional technology.

2. The present invention can correct the innately versatile address to a more complete and valid address to compensate for the incorrect filling of the free-form address.

3. The present invention can reduce the bounce rate to avoid excess postage costs and increase the effective mailing ratio.

4. The invention can ensure the correctness and integrity of the address information after the transfer, prevent the occurrence of the event that the letter cannot be delivered, and protect the rights of the user and the merchant.

As shown in Figure 1, the implementation method of the design method of the present invention can be divided into a Chinese address keyword fast search program 1 and a conversion rule correction program 2, through the postal code: 82442, address: 360 Anxin Road, Jiaosu Village, Yancao District, Gao City As an example, the description is as follows.

Referring to FIG. 2, the Chinese address keyword fast search program 1 is composed of a search index component 11, a boundary value detecting component 12, a keyword checking unit 13, a region level checking unit 14, and an address cutting component 15; The address in the free format can be correctly cut out. This program uses the keyword position probability statistics to cut; firstly, the keywords that should appear in the region level are determined for different regional levels, and the regional level is divided by the keyword level. For example, the county and the city are at the same regional level, and the village and the middle are at the same regional level, but the county and the village are different regional levels, and the different levels have a sequential relationship with Keyword(t,i). Define the ith keyword that will appear in the tth region hierarchy. The keyword position probability value can be counted through the target database or with a known database. The target database can be a storage space for storing the pending data. The known database can be provided by a third party, such as a database of post office addresses. The target database has at least a postal code number and an address. The known database field is not limited to the postal area code, address, and regional level. The known database is used to know the keywords included in each level, and the target can be accessed through the target. As a supplement, the database counts the probability value of the address string position of each address level in the address field, and the search index component 11 indicates the position of the area level keyword by X(t, n, i). The probability value appears as follows:

The X(t, n, i) is output to the boundary value detecting component 12, and the boundary value at which the nth position has the highest probability is searched for. According to the probability value of each Keyword(t, i) at the nth position, Indicates the most frequent location of the keyword, as shown in equation (2): Will be the largest boundary position Output to the keyword checking unit 13, the boundary position is the length of each address area, and it is judged whether the character of the boundary value position is a keyword (Boundary_keyword), and the checking method is as follows: Boundary_keyword Keyword(t,i)

If the character at the position of the boundary value is not the keyword of the same level, the current boundary value position can be removed, and the boundary value detecting component 12 can be re-executed to find the position of the probability of occurrence of the second largest boundary value; When the keyword is reached, the keyword of the next level will be searched. Once the keyword checking unit 13 judges the keyword, the keyword checking unit 14 is entered to ensure that each of the determined regional levels is checked. The maximum area level value is accumulated by the area level parameter to execute the boundary value detecting element 12 of the next area level. Finally, when each of the determined area levels finds the boundary value position, the processing of the address cutting element 15 is performed. The program defines the output of this component as result _t , which indicates the cutting result of the t-th region level, and completes the Chinese address keyword fast search procedure 1 .

The conversion rule correction program 2 can perform the above-mentioned cutting result according to different purposes, and convert the cut string into a correct string as one of the processing methods of the program. Because the free-form postal area code field has a relatively high correct ratio compared to the address field, the conversion rule correction program 2 corrects the output of the Chinese address keyword quick search program 1 to the correct result through the postal area code, and then The zip code, address and area level field of the known database are represented by Addr _ String ( t , zipcode ) to retrieve the string of the t-th region level by using the zip code, and the Addr _ String ( t , zipcode ) can be used by the target database. Or a database provided by a known database, and the result _{t is} converted to a string of Addr _ String ( t , zipcode ).

Take the above-mentioned user input zip code: 82442 and address: No. 360, Anxin Road, Jiaosu Village, Yancao District, Gaocheng City. For example, in the Chinese address keyword quick search procedure 1, assume the region level to the fourth layer to be found. . In the first regional level, the keywords are “County” and “City”. The maximum probability position of the keyword is the third position, and the extracted keyword is “smoke”. The result of entering the keyword check unit is found. "Smoke" is not the first layer of keywords, the algorithm will remove the third position, to the boundary value detection component 12 to find the next largest probability position for the second position, according to the second position of the address The keyword is "City", which matches the keyword of the first regional level, so the boundary value of the first regional level is two, then the regional level parameters are accumulated, and so on, to find the boundary value of the next regional level. In the address cutting element 15, the boundary value of each region level is cut, so that the cutting results of "High City", "Yancao District", "Kuansu Village" and "Anxin Road" are obtained, and the conversion rule is corrected. 2 program by the zip code as a secondary correction, in which the zip code is 82442, out Addr _ String (1,82442) = Kaohsiung, Addr _ String (2,82442) = yanchao district, the correction result is "Kaohsiung""Yanchao District , "Spica Village" and "On the new road."

The detailed description of the present invention is intended to be illustrative of a preferred embodiment of the invention, and is not intended to limit the scope of the invention. The patent scope of this case.

To sum up, this case is not only innovative in terms of technical thinking, but also has many of the above-mentioned functions that are not in the traditional methods of the past. It has fully complied with the statutory invention patent requirements of novelty and progressiveness, and applied for it according to law. Approved this invention patent application, in order to invent invention, to the sense of virtue.

1‧‧‧Chinese address keyword quick search program

2‧‧‧ Conversion Rule Amendment Procedure

11‧‧‧Search index component

12‧‧‧Boundary value detection component

13‧‧‧Keyword checking unit

14‧‧‧Regional level inspection unit

15‧‧‧ Address Cutting Components

The detailed description of the present invention and the accompanying drawings will be further understood, and the technical contents of the present invention and the functions thereof can be further understood. FIG. 1 is a flow chart of the Chinese address translation method of the hierarchical architecture of the present invention. Figure 2 is a fast search program diagram of Chinese address keywords for the Chinese address translation method of the hierarchical architecture of the present invention.

1‧‧‧Chinese address keyword quick search program

2‧‧‧ Conversion Rule Amendment Procedure

Claims (10)

A Chinese-language address translation method for hierarchical architecture, the data source is a free format, and stores a string with continuous and hierarchical relationship, including the following steps: a. first setting a keyword of the current regional level according to a statistical method a probability value appearing at each position in the string; b. calculating a sum of the probability values of each of the keywords at the current level of the region at each position, if the higher position character in the sum of the probability values is included in the current region The keyword of the level, which is the boundary value of the current level of the region, otherwise it will continue to search for the location where the probability value is the second highest and is included in the current level of the region; c. Finally, according to the region level The boundary value is cut to obtain a string of each of the region levels. For example, the Chinese address translation method of the hierarchical structure described in claim 1 is that the regional level is a keyword having a hierarchical relationship. The Chinese address translation method of the hierarchical structure described in claim 1, wherein the establishment of the regional level and the keyword is statically set or dynamically changed. For example, the Chinese address translation method of the hierarchical structure described in claim 1 is the target database or the known database. For example, the Chinese address translation method of the hierarchical structure described in the fourth application of the patent scope is a storage space for storing the data to be processed. The Chinese address of the hierarchical structure as described in item 4 of the patent application scope In other words, the known database is provided by a third party. The Chinese address translation method of the hierarchical structure as described in claim 6 wherein the third party provides a database of post office addresses. For the Chinese address translation method of the hierarchical structure described in claim 4, the field of the target database includes the zip code and the address. For the Chinese address translation method of the hierarchical architecture described in claim 4, the fields in the known database are not limited to the postal area code, the address and the regional level. For example, the Chinese address translation method of the hierarchical structure described in claim 1 further includes processing the output layer level string and correcting the string to the correct string value, and the correct The string value is provided by the target database or the known database.