CN114625845A

CN114625845A - Information retrieval method, intelligent terminal and computer readable storage medium

Info

Publication number: CN114625845A
Application number: CN202011444290.0A
Authority: CN
Inventors: 王妍
Original assignee: Shenzhen TCL New Technology Co Ltd
Current assignee: Shenzhen TCL New Technology Co Ltd
Priority date: 2020-12-11
Filing date: 2020-12-11
Publication date: 2022-06-14

Abstract

The invention discloses an information retrieval method, an intelligent terminal and a computer readable storage medium, wherein the method comprises the following steps: acquiring an instruction sentence sent by a user; determining a replacement word corresponding to the instruction statement according to a target word in the instruction statement; and determining and outputting target information corresponding to the instruction sentence according to the alternative words. The invention can improve the accuracy of the information retrieval result.

Description

Information retrieval method, intelligent terminal and computer readable storage medium

Technical Field

The present invention relates to the field of computer technologies, and in particular, to an information retrieval method, an intelligent terminal, and a computer-readable storage medium.

Background

With the development of natural language processing technology, users can issue various types of instructions through voice or text, such as controlling power on and off, opening a specific software, and the like. However, such instructions are generally few and fixed, for example, a user wants to turn on or off the computer, the instructions are generally "turn on" or "turn off", the software a is opened, and the instructions are generally "open the software a".

However, when information retrieval is performed through voice or text, an instruction issued by a user is often encountered, and a terminal can only search related information according to a target word in the instruction of the user, so that the obtained information is coarse and difficult to meet the requirements of the user. For example, a user wants to listen to a song of a certain band, and the issued instruction is "find a song of the certain band", the terminal may normally identify, but the band may have an abbreviation, a transliterated name, an ideographic name, and an love, and different titles may appear at different times. Therefore, when the instruction issued by the user is an instruction related to retrieval, i.e. an instruction sentence, the terminal often cannot accurately identify and provide an effective retrieval result.

Disclosure of Invention

The invention mainly aims to provide an information retrieval method, an intelligent terminal and a computer readable storage medium, and aims to solve the problem that an effective retrieval result cannot be provided according to instruction sentences of a user in the prior art.

In order to achieve the above object, the present invention provides an information retrieval method, including the steps of:

acquiring an instruction sentence sent by a user;

determining a replacement word corresponding to the instruction statement according to a target word in the instruction statement;

and determining and outputting target information corresponding to the instruction sentence according to the alternative words.

Optionally, the information retrieval method, where the determining, according to the target word in the instruction statement, the replacement word corresponding to the instruction statement specifically includes:

determining a target word corresponding to the instruction sentence according to a preset target word bank;

and determining a replacement word corresponding to each target word according to a preset replacement word library.

Optionally, the information retrieval method, where the determining, according to a preset target word library, a target word corresponding to the instruction sentence specifically includes:

performing word segmentation on the instruction sentence to generate a plurality of text character strings;

and determining the target words in the text character strings according to the similarity values of the text character strings and the target words of all the keywords in a preset target word bank.

Optionally, in the information retrieval method, the replacement word includes a hypernym, a hyponym, a synonym, and a hyponym corresponding to a keyword in the target thesaurus; before determining the replacement word corresponding to each target word according to a preset replacement word bank, the method further includes:

and aiming at each keyword, taking the hypernym, the hyponym, the synonym and the synonym corresponding to the keyword as corresponding alternative words according to a preset knowledge base.

Optionally, in the information retrieval method, the replacement word includes an alias corresponding to a keyword in the target thesaurus; before determining the replacement word corresponding to each target word according to a preset replacement word bank, the method further includes:

acquiring and collecting social sentences, clustering the social sentences to generate a plurality of social sentence groups;

determining reference sentences in the social sentence groups according to preset reference sentence rules;

aiming at each social sentence group, determining an alias corresponding to a reference character string in a reference sentence in each social sentence in the social sentence group according to the reference sentence in the social sentence group, wherein the reference character string is a character string corresponding to a keyword in the keyword library;

and determining the alternative name corresponding to the keyword according to the corresponding relation between the reference character string and the keyword, and taking the alternative name as a corresponding replacement word.

Optionally, the information retrieval method, where the determining, according to a preset reference statement rule, a reference statement in each social statement group specifically includes:

calculating sentence similarity values among social sentences in each social sentence group aiming at each social sentence group;

and determining a reference sentence in the social sentence group according to the sentence similarity value.

Optionally, the information retrieval method, where the determining and outputting the target information corresponding to the instruction statement according to the replacement word specifically includes:

combining the replacement words corresponding to the target words to generate a plurality of replacement word sets;

and taking the target words as a target word set, and determining and outputting target information corresponding to the instruction sentences according to the replacement word set and the target word set.

Optionally, the information retrieval method, wherein the determining and outputting the target information corresponding to the instruction statement according to the replacement word set and the target word set specifically includes:

for each replacement word set, determining corresponding data information according to the replacement words in the replacement word set; and

determining corresponding data information according to the target words in the target word set;

and sequencing the data information according to a preset sequencing rule, generating and outputting target information corresponding to the instruction statement.

In addition, to achieve the above object, the present invention further provides an intelligent terminal, wherein the intelligent terminal includes: a memory, a processor and an information retrieval program stored on the memory and executable on the processor, the information retrieval program when executed by the processor implementing the steps of the information retrieval method as described above.

Further, to achieve the above object, the present invention also provides a computer-readable storage medium storing an information retrieval program which, when executed by a processor, implements the steps of the information retrieval method as described above.

After the instruction sentence is obtained, the information retrieval is not conventionally carried out according to the target word playing the role of understanding the meaning in the instruction sentence, but the possible alternative words are determined according to the target word in the instruction sentence, and then the corresponding target information is searched according to the target word and the alternative words. Since the replacement of the instruction sentence by the replacement word does not make ambiguity on understanding the instruction sentence, but can enlarge the searched object, thereby providing more accurate search results.

Drawings

FIG. 1 is a flow chart of a preferred embodiment of the information retrieval method of the present invention;

FIG. 2 is a flowchart of step S200 according to the preferred embodiment of the present invention;

FIG. 3 is a flowchart of step S210 in the preferred embodiment of the information retrieval method of the present invention;

fig. 4 is a schematic operating environment diagram of an intelligent terminal according to a preferred embodiment of the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

As shown in fig. 1, the information retrieval method according to the preferred embodiment of the present invention includes the following steps:

and step S100, acquiring an instruction sentence sent by a user.

Specifically, in the present embodiment, the main body executing the information retrieval method is helper software installed in the intelligent terminal. When the user uses the intelligent terminal, the user can give an instruction through an entity keyboard, a virtual keyboard or a microphone of the intelligent terminal. Such as "find the latest album of the a band". When the instruction sentence is in a voice form, the intelligent terminal needs to adopt voice recognition to convert the instruction sentence into a text-form instruction sentence and store the text-form instruction sentence into the local for convenient processing after acquiring the instruction sentence of the user through the microphone. The helper software then retrieves the instruction statement locally. The assistant software can be in a real-time running state or a sleeping state, and if the assistant software is in the sleeping state, the assistant software can be awakened to work after the intelligent terminal acquires the instruction statement.

And step S200, determining a replacement word corresponding to the instruction statement according to the target word in the instruction statement.

Specifically, after the instruction sentence is obtained, the helper software determines the target word corresponding to the instruction sentence. The target word refers to a word that plays a key role in understanding the instruction sentence. A replacement word bank is preset, replacement words possibly corresponding to each word are stored in the replacement word bank, and the replacement words refer to words which replace corresponding words in the instruction sentences and do not affect the meaning of the whole instruction sentences. Because the replacement words and the words in the replacement word library correspond to each other, the replacement words corresponding to the instruction sentence can be determined according to the corresponding relation and the target words in the instruction sentence.

Further, referring to fig. 2, step S200 includes:

step S210, determining a target word in the instruction sentence according to a preset target word library.

Specifically, a target word stock is set in advance, and a plurality of keywords are stored in the target word stock. In a first implementation manner of this embodiment, the process of determining the target word is: and traversing sequentially according to the keywords of the target word bank and the sentence sequence in the instruction sentence, and determining that the keywords are the target words in the instruction sentence when the keywords in the instruction sentence are completely matched with the keywords in the target word bank.

Further, since the division between the chinese words is not determined according to the sentence order, for example, "there is a battery in the office," according to the sentence order, it should be divided into "office," "indoor," "powered" and "pool," but this completely does not conform to the true meaning of the sentence, so there is a big deficiency in adopting the first implementation manner, referring to fig. 3, in this embodiment, in order to improve the accuracy of the target word retrieval, the method of determining the target word is:

step S211, performing word segmentation on the instruction sentence, and generating a plurality of text character strings.

Specifically, the instruction sentence is first participled to generate a plurality of text strings. Word segmentation refers to the splitting of text into the smallest units of phonetic representation. The present embodiment is preferably performed by using a statistical-based word segmentation method, such as a conditional random field-based word segmentation method, a hidden markov model-based word segmentation method, and a deep learning-based word segmentation method. With the above word segmentation prevention, the instruction sentence is split into a plurality of text strings, for example, "find the latest album of the a band" is split into "find", "a band", "of", "latest", and "album".

Step S212, determining a target word in the text character string according to the similarity value of the text character string and the target word of each keyword in a preset target word bank.

Specifically, the similarity value of the target word is calculated by using the obtained text character string and each keyword in a preset target word library. For example, common keyword similarity calculation based on word vectors converts text character strings and keywords in a target word library into a vector form in a word2vec manner, and the like, and because the vector has a numerical value and a direction, the difference between the two can be calculated. The difference describing method generally adopts word shift distance and cosine value, and the larger the word shift distance is, the farther the word shift distance is, the longer the distance between the word shift distance and the cosine value is, so the similarity value is smaller; the smaller the word shift distance, the closer the word shift distance and the word shift distance, and thus the larger the similarity value. The cosine value is the cosine value of the included angle of two vectors to describe the similarity of the two vectors. The closer the cosine value is to 1, the more only 0 degrees of the angle between the two vectors, the more similar the vectors and hence the smaller the similarity value.

In this embodiment, since each word in the instruction sentence is not necessarily a target word, for example, "in the example sentence," a target word similarity value threshold is set in advance, then, for each text character string, a target word similarity value between the text character string and each keyword in the target word library is calculated, and then, the text character string corresponding to the target word similarity value which exceeds the target word similarity threshold and has the largest numerical value is selected as the target word.

Step S220, determining a replacement word corresponding to each target word according to a preset replacement word library.

Specifically, a replacement word bank is preset, where the replacement word bank includes a replacement word corresponding to each keyword in the target word bank, and the replacement word refers to a keyword that can replace the keyword in a sentence and does not affect the meaning of the sentence. For example, the replacement word of the "search" may be a word such as "search", "retrieve", and the like, and the replacement word is associated with each keyword in the target word bank. Because the target words are determined through the target word bank and the corresponding keywords of the target words exist in the target word bank, after the target words are determined, the corresponding replacement words are searched in the replacement word bank according to the target words.

Further, the replacement words comprise hypernyms, hyponyms, synonyms and synonyms corresponding to the keywords in the target word bank; before step S220, the method further includes: and aiming at each keyword, taking the hypernym, the hyponym, the synonym and the synonym corresponding to the keyword as corresponding replacement words according to a preset knowledge base.

Specifically, the context is a linguistic concept. The more general words are called the superior words of the more specific words, and the more specific words are called the inferior words (hyponyms) of the more general words. Synonyms refer to keywords having relatively similar meanings, while synonyms refer to two words having the same meaning. For example, the hypernym of "video" includes "audio", the hyponym of audio includes "video", the synonyms and synonyms of "find" includes "find", and the like. In the embodiment, the established knowledge base comprises the relationships among the known superior words, inferior words, synonyms and similar words, and forms a network of the superior and inferior keywords and similar words related to the keywords. And then determining the superior word, the inferior word, the synonym and the near-meaning word corresponding to the keyword in the target word bank according to the relation between the words in the established knowledge bank, and taking the words as the replacement words corresponding to the keyword in the target word bank and storing the words in the replacement word bank.

Further, the replacement word comprises an alias corresponding to the target word; before step S220, the method further includes:

step A10, acquiring and collecting social sentences, and clustering the social sentences to generate a plurality of social sentence groups.

Specifically, since hypernyms, hyponyms, synonyms, and hypernyms are obtained based on relationships between given words, as networks develop, the meanings of many words are rapidly developing, and changes in the meanings of many words are generated based on social networks. For example, "a band" is abbreviated as "B" and the transliteration name is "C", but "B" and "C" are both equivalent to "a band". Therefore, a large number of social statements are collected first.

And after the social sentences are collected, clustering the social sentences. In the preferred clustering method in this embodiment, a plurality of social sentence groups based on the social field are obtained by classifying according to the field corresponding to the social sentence based on supervised clustering, and then supervised or unsupervised clustering is performed on the social sentences in the social sentence groups based on the social field. The purpose of distinguishing between social domains is because the meaning of the same word may vary in different domains. For example, "compendium" refers to a summary or detailed rule in the literature domain, and two grades of bio-classification based on the forest tolerant bio-classification method in the biological domain. Therefore, the domain corresponding to the social statement is divided first. The implementation of social domain partitioning may be based on an acquisition platform implementation of social statements, such as medically related, astronomical related, entertainment related, and the like. And after the social sentences are obtained, grouping the social sentences in the same social field by adopting a supervised or unsupervised clustering mode to obtain a plurality of social sentence groups.

Step A20, determining the reference sentence in each social sentence cluster according to the preset reference sentence rule.

Specifically, the reference sentence refers to a sentence included in the social sentence group and corresponding to the keyword in the keyword library. Since social sentences in the social sentence group are associated with each other, if a certain social sentence contains a keyword, an alias corresponding to the social sentence, that is, a reference sentence, is likely to appear in similar social sentences. Therefore, the alias corresponding to the keyword may be determined to appear in other social sentences in the social sentence group according to the reference sentence.

Further, step a20 includes:

in step A21, for each social sentence group, a sentence similarity value between the social sentences in the social sentence group is calculated.

Specifically, the sentence similarity value between the social sentences in the same social sentence group is calculated by taking each social sentence group as a unit. The way of calculating the similarity of sentences is similar to the way of calculating the similarity of keywords, except that the vectors obtained by words are generally small and the calculation is fast, and the sentences are long, so the vectors generally exist in a matrix form and the calculation is complex. The calculation method includes calculation of the jaccard coefficient, calculation of the edit distance, etc., which are not described herein again.

Step A22, determining the reference sentence in the social sentence group according to the sentence similarity value.

Specifically, the sizes of the sentence similarity values are compared, and then the social sentences in the social sentence group with higher social sentence similarity values are selected as the reference sentences.

It is noted that a social sentence group may include a plurality of social sentence subgroups depending on the degree of detail of the grouping. Taking the medical field as an example, the social sentence group based on medical science can be further divided into traditional Chinese medicine and western medicine, the western medicine can be divided into clinical medicine and basic medicine, and the basic medicine can be further divided into medical biochemistry, human immunology and the like. The lowest level of social sentence group is encompassed by the previous level of social sentence group. And determining the reference sentence in the social sentence group is generally to select the determination in the social sentence group of the lowest level.

Step a30, for each social sentence group, determining an alias corresponding to a reference character string in a reference sentence in each social sentence in the social sentence group according to the reference sentence in the social sentence group, where the reference character string is a character string corresponding to the keyword.

Specifically, after determining the reference sentence in each social sentence group, determining a character string in the reference sentence, which is equal to the keyword, as a reference character string according to the keyword in the keyword library in the manner of determining the target word in the instruction sentence. Then, for each social statement, performing word segmentation on the social statement to obtain a plurality of character strings, then calculating alias similarity values between the character strings and reference character strings, and then taking the character strings with the alias similarity values exceeding a preset alias similarity threshold value as aliases corresponding to the reference characters.

Step A40, determining the alternative name corresponding to the keyword according to the corresponding relationship between the reference character string and the keyword, and using the alternative name as a corresponding alternative word.

Specifically, since the reference character string and the keyword are in a corresponding relationship, after the alias corresponding to the reference character string is determined, the alias corresponding to the keyword can be determined according to the corresponding relationship between the alias and the keyword, and the alias is used as the replacement word corresponding to the keyword.

And step S300, determining and outputting target information corresponding to the instruction sentence according to the replacement words and the target words.

Specifically, after determining the replacement words corresponding to the keywords, for example, the replacement words corresponding to the above-mentioned "a band" include "B" and "C". And then, according to the replacement words and the target words, performing data retrieval, thereby determining and outputting target information corresponding to the instruction sentence.

Further, step S300 includes:

step S310, combining the replacement words corresponding to each target word to generate a plurality of replacement word sets.

Specifically, taking target words "find", "a band", "latest", and "album" as examples, the substitute word for "find" is "find", "search", etc., the substitute word for "a band" is "B" and "C", etc., and the substitute word for "latest" is "latest", and the substitute word for "album" is "large disc", "music collection". Randomly selecting one of the replacement words corresponding to the search, such as the search, then randomly selecting one of the replacement words corresponding to the A band, such as the B, then selecting the nearest and the large saucer as candidate words, and then combining the search, the B, the nearest and the large saucer to generate a replacement word set. Since the replacement words selected for combination are different and the replacement words in the replacement word set are different, a plurality of replacement word sets are generated.

And step S320, taking the target words as a target word set, and determining and outputting target information corresponding to the instruction sentences according to the replacement word set and the target word set.

Specifically, after a plurality of replacement word sets are obtained, the target words are used as a group of target word sets, and according to the target word sets and the replacement word sets, data information corresponding to the target word sets and the replacement word sets is obtained in a pre-connected database and is used as target information corresponding to the instruction sentences.

Further, since there is much data information obtained from the set of replacement words and the set of target words, step S320 includes:

step S321, aiming at each replacement word set, determining corresponding data information according to the replacement words in the replacement word set; and determining corresponding data information according to the target words in the target word set.

Specifically, a target word in the target word set is used as a search word, and the search is performed in a database connected in advance, so that data information corresponding to the target word set is determined. Meanwhile, aiming at each alternative word set, the alternative words in the alternative word set are used as search words, and the search is carried out in a database connected in advance, so that the data information corresponding to the alternative word set is determined. Finally, a plurality of pieces of relevant data information are obtained.

Step S322, according to a preset sorting rule, sorting the data information, and generating target information corresponding to the instruction statement.

Specifically, the sorting rule is a rule for sorting the data information, and may be a rule for calculating similarity values between the data information and the instruction statements and sorting the data information according to the similarity values; different weighting values given to different keywords or alternative words in advance can also be adopted, for example, if a certain alias has the highest use frequency and the widest use range, the weighting of the alias is increased. After the data information is obtained, the weight value corresponding to the data information is calculated according to the weight value of the replacement word or the keyword corresponding to the data information, and then the data information is sequenced according to the weight value to generate target information corresponding to the instruction statement.

Further, after the target information is obtained, the corresponding instruction is executed according to the target information. For example, the target information obtained is "song list in the latest album of the a band" described above, and this album is played according to the song list. Therefore, the scheme can be combined with the execution instruction, and the corresponding operations are sequentially executed according to the target information.

Further, as shown in fig. 4, based on the above information retrieval method, the present invention also provides an intelligent terminal, which includes a processor 10, a memory 20 and a display 30. Fig. 4 shows only some of the components of the smart terminal, but it should be understood that not all of the shown components are required to be implemented, and that more or fewer components may be implemented instead.

The memory 20 may be an internal storage unit of the intelligent terminal in some embodiments, such as a hard disk or a memory of the intelligent terminal. The memory 20 may also be an external storage device of the Smart terminal in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the Smart terminal. Further, the memory 20 may also include both an internal storage unit and an external storage device of the smart terminal. The memory 20 is used for storing application software installed in the intelligent terminal and various data, such as program codes of the installed intelligent terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In an embodiment, the memory 20 stores an information retrieval program 40, and the information retrieval program 40 can be executed by the processor 10, so as to implement the information retrieval method in the present application.

The processor 10 may be a Central Processing Unit (CPU), microprocessor or other data Processing chip in some embodiments, and is used for executing program codes stored in the memory 20 or Processing data, such as executing the information retrieval method.

The display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch panel, or the like in some embodiments. The display 30 is used for displaying information at the intelligent terminal and for displaying a visual user interface. The components 10-30 of the intelligent terminal communicate with each other via a system bus.

In one embodiment, when the processor 10 executes the information retrieval program 40 in the memory 20, the following steps are implemented:

acquiring an instruction sentence sent by a user;

Wherein, the determining, according to the target word in the instruction statement, the replacement word corresponding to the instruction statement specifically includes:

The determining, according to a preset target word bank, a target word corresponding to the instruction sentence specifically includes:

The replacement words comprise hypernyms, hyponyms, synonyms and synonyms corresponding to the keywords in the target word bank; before determining the replacement word corresponding to each target word according to a preset replacement word bank, the method further includes:

and aiming at each keyword, taking the hypernym, the hyponym, the synonym and the synonym corresponding to the keyword as corresponding replacement words according to a preset knowledge base.

Wherein the replacement words comprise aliases corresponding to the keywords in the target word bank; before determining the replacement word corresponding to each target word according to a preset replacement word bank, the method further includes:

acquiring and collecting social sentences, and clustering the social sentences to generate a plurality of social sentence groups;

Wherein, according to a preset reference sentence rule, determining a reference sentence in each social sentence cluster specifically includes:

calculating sentence similarity values among the social sentences in each social sentence group aiming at each social sentence group;

Determining and outputting target information corresponding to the instruction statement according to the alternative word, specifically including:

Wherein, the determining and outputting the target information corresponding to the instruction sentence according to the replacement word set and the target word set specifically includes:

The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores an information retrieval program, which when executed by a processor implements the steps of the information retrieval method as described above.

Of course, it will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program, and the program can be stored in a computer readable storage medium, and when executed, the program can include the processes of the embodiments of the methods described above. The computer readable storage medium may be a memory, a magnetic disk, an optical disk, etc.

It is to be understood that the invention is not limited to the examples described above, but that modifications and variations may be effected thereto by those of ordinary skill in the art in light of the foregoing description, and that all such modifications and variations are intended to be within the scope of the invention as defined by the appended claims.

Claims

1. An information retrieval method, characterized by comprising:

acquiring an instruction sentence sent by a user;

2. The information retrieval method according to claim 1, wherein the determining, according to the target word in the instruction sentence, the replacement word corresponding to the instruction sentence specifically includes:

3. The information retrieval method according to claim 2, wherein the determining the target word corresponding to the instruction sentence according to a preset target word bank specifically includes:

4. The information retrieval method according to claim 2, wherein the replacement word includes an hypernym, a hyponym, a synonym, and a homonym corresponding to the keyword in the target thesaurus; before determining the replacement word corresponding to each target word according to a preset replacement word bank, the method further includes:

5. The information retrieval method according to claim 2, wherein the replacement word includes an alias corresponding to a keyword in the target thesaurus; before determining the replacement word corresponding to each target word according to a preset replacement word bank, the method further includes:

aiming at each social statement group, determining an alias corresponding to a reference character string in a reference statement in each social statement in the social statement group according to the reference statement in the social statement group, wherein the reference character string is a character string corresponding to a keyword in the keyword library;

6. The information retrieval method according to claim 5, wherein the determining a reference sentence in each social sentence cluster according to a preset reference sentence rule specifically includes:

7. The information retrieval method according to any one of claims 1 to 6, wherein the determining and outputting target information corresponding to the instruction sentence according to the replacement word specifically includes:

8. The information retrieval method according to claim 7, wherein the determining and outputting the target information corresponding to the instruction sentence according to the replacement word set and the target word set specifically includes:

9. An intelligent terminal, characterized in that, intelligent terminal includes: memory, processor and an information retrieval program stored on the memory and executable on the processor, the information retrieval program when executed by the processor implementing the steps of the information retrieval method as claimed in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that the computer-readable storage medium stores an information retrieval program, which when executed by a processor implements the steps of the information retrieval method according to any one of claims 1 to 8.