Search device, search program
The search device improves product search accuracy by inserting delimiters and handling kana characters, effectively extracting desired products by addressing issues with missing character strings and consecutive identical characters in search keywords.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- HAYASHI KK
- Filing Date
- 2024-11-07
- Publication Date
- 2026-05-19
AI Technical Summary
Users often fail to find desired products in search systems due to missing middle character strings or incorrect handling of kana characters and consecutive identical characters in search keywords, leading to inappropriate extraction of products.
A search device that inserts delimiters into input strings to generate search strings, handles kana characters, and manages consecutive identical characters to improve product extraction accuracy.
Enhances product search efficiency by accurately extracting desired products, reducing the need for re-entry or correction of input strings, especially in voice input scenarios.
Smart Images
Figure 2026082343000001_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to a search device and a search program for searching for a desired product from products registered in a database.
Background Art
[0002] Conventionally, a technology of a search device that inputs a search keyword such as a product name into a search box of an e-commerce site and extracts a desired product is known. For example, Patent Document 1 introduces an invention of a product presentation system that presents a product related to the product together with a product that matches the search keyword.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0004] By the way, a user does not always remember the official name of a product. For this reason, for example, when wanting to purchase a product "welding 90L long elbow", if a search keyword such as "welding long elbow" is input into the search box, there is a problem that the product is not appropriately extracted because the middle character string "90L" of the product is missing.
[0005] Therefore, an object of the present disclosure is to provide a search device and a search program that can easily search for products from a database.
Means for Solving the Problems
[0006] To solve the above problems, the search device of this disclosure is a search device for searching for products registered in a database, comprising: an input string acquisition means for acquiring an input string entered by a user; a search string generation means for generating a search string based on the input string; and an extraction means for extracting products from the database using the search string, wherein the search string generation means generates the search string by inserting a delimiter into the input string. [Effects of the Invention]
[0007] According to the search device disclosed herein, a delimiter is inserted into the input string entered by the user to generate a search string, thereby enabling the extraction of products containing the string that makes up the input string. As a result, products are more likely to be found, and products can be easily searched. [Brief explanation of the drawing]
[0008] [Figure 1] This is a schematic diagram of a search device showing one embodiment of the present disclosure. [Figure 2] This is a block diagram showing the configuration of the search device. [Figure 3] (a) The conventional search string, and (b) the extraction result when using the search string of this disclosure. [Figure 4] This is an explanatory diagram illustrating the handling of kana characters that are the same as particles. [Figure 5] This diagram illustrates the extraction results when (a) a conventional search string and (b) a search string of the present disclosure are used after deleting kana characters identical to particles. [Figure 6] This is an explanatory diagram illustrating how to handle consecutive identical characters (numbers). [Figure 7] This is an explanatory diagram illustrating how to handle consecutive identical characters (Kanji). [Figure 8] This is a flowchart showing the search process. [Modes for carrying out the invention]
[0009] Hereinafter, an embodiment of the present disclosure implemented in the search device 1 will be described with reference to the drawings. In the following description, the delimiter character (space s) in the search string k will be represented as "□".
[0010] As shown in Figure 1, the search device 1 searches for products registered in database 3. The search device 1 is installed on e-commerce sites and apps, and when a user purchases a product via the internet N, they can enter search keywords into the e-commerce site or app to search for the desired product. The search program is executed on information terminal 2, causing information terminal 2 to function as search device 1. Database 3 is located on a server 4, such as a product vendor's server or a cloud server.
[0011] The monitor 21 of the information terminal 2 displays a search box 22 and a search button 23. When the user enters a search keyword into the search box 22 and taps the search button 23, a list of search results 24 is displayed.
[0012] As shown in Figure 2, the search device 1 includes an input string acquisition unit 11 that acquires an input string n entered by the user, a search string generation unit 13 that generates a search string k based on the input string n, and an extraction unit 14 that extracts products from the database 3 using the search string k.
[0013] The search device 1 may also be equipped with an extraction trial unit 12 that attempts to extract products from the database using an input string n. In this case, if the extraction trial unit 12 extracts a product, the search device 1 omits the processing of the search string generation unit 13 and the extraction unit 14. On the other hand, if the extraction trial unit 12 does not extract a product, the search string generation unit 13 generates a search string k, and the extraction unit 14 extracts the product. In other words, if the search keyword entered by the user is appropriate, the search device 1 can omit subsequent processing, thereby reducing the processing burden on the search device 1.
[0014] The search string generation unit 13 generates a search string k by inserting a delimiter character (space s) into the input string n. Specifically, the search string generation unit 13 inserts a space s for each character in the input string n to generate the search string k. Also, when the characters before and after the position where the space s is to be inserted are spaces s, the search string generation unit 13 does not insert the space s. Here, the "delimiter character" functions as "AND" in the query to the database 3. For example, "string a" AND "string b" constitutes a query to "extract products containing string a and string b". In the present disclosure, a space s is adopted as the delimiter character, but other delimiter characters can also be used. Hereinafter, a specific example will be given to explain the search string k generated by the search string generation unit 13 and the extraction result when the search string k is used for extraction.
[0015] Fig. 3 shows an example where the desired product is a product containing "welding" and "long elbow", and the input string n is "welding long elbow".
[0016] As shown in Fig. 3(a), when extracting using the conventional search string k' "welding long elbow", products containing "90L" between "welding" and "long elbow" cannot be extracted.
[0017] On the other hand, as shown in Fig. 3(b), when a space s is inserted into the input string n, the search string k becomes "w □ e □ l □ d □ i □ n □ g □ l □ o □ n □ g □ e □ l □ b □ o". At this time, the extraction unit 14 extracts products containing "w", "e", "l", "d", "i", "n", "g", "l", "o", "n", "g", "e", "l", and "b" from the database 3. As a result, desired products containing each of these characters, such as "welding 90L long elbow SUS304 S / 40 - 15", are extracted.
[0018] When the input string n contains kana characters identical to particles, the search string generation unit 13 deletes the kana characters. Examples of kana characters identical to particles include, for example, "の" and "ノ". These days, the cases of using voice input functions such as smartphones are increasing. In voice input, for example, in the case of "40 of the 90L long elbow for welding", the particle "の" is often inserted between the product name and the size, and as a result, the product is often not extracted. By deleting the kana characters identical to particles, products can be appropriately extracted. Hereinafter, specific examples will be given for explanation.
[0019] FIG. 4 shows an example in which the desired product is a product containing "welding", "long elbow", and "40", and the input string n is "welding long elbow の 40".
[0020] As shown in FIG. 4(a), when the kana character "の" identical to the particle is not deleted, the search string k becomes "溶□接□ロ□ン□グ□エ□ル□ボ□の□4□0". Since there is no product containing "溶" and "の", the product is not extracted.
[0021] On the other hand, as shown in FIG. 4(b), when the kana character "の" identical to the particle is deleted, the search string k becomes "溶□接□ロ□ン□グ□エ□ル□ボ□4□0". In this case, by the same operation as in the case of FIG. 3(b), products such as "welding 90L long elbow SUS304 S / 40-15" are extracted.
[0022] The method of deleting the kana characters identical to particles functions appropriately even when the product name contains kana characters identical to particles. Hereinafter, specific examples will be given for explanation.
[0023] FIG. 5 shows an example in which the desired product is "PVC processed bamboo shoot", and the input string n is "PVC processed bamboo shoot".
[0024] As shown in FIG. 5(a), when extracted using the conventional search string k' "PVC processed bamboo shoot", since there is a product containing the entire input string n, the desired product "PVC processed bamboo shoot -25" etc. are extracted.
[0025] As shown in FIG. 5(b), when deleting the kana character "no" that is the same as the particle, the search string k becomes "salt □ vinyl □ product □ processing □ bamboo □ child". Also in this case, products that include all of "salt", "vinyl", "product", "processing", "bamboo", and "child" are extracted, and "salt vinyl processed bamboo child -25" etc. are extracted. Thus, the method of deleting the kana character that is the same as the particle functions without problems even when the product name includes a kana character that is the same as the particle.
[0026] When the input string n includes a plurality of consecutive same-type characters, it is preferable that the search string generation unit 13 does not insert a space s between the same-type characters. Here, "same-type characters" refer to cases where the characters are all numbers, all Chinese characters, all English letters, all hiragana, or all katakana. English uppercase and lowercase letters are not distinguished. Full-width and half-width characters are not distinguished. Whether a plurality of consecutive characters are the same-type characters can be determined using the character code. A plurality of consecutive same-type characters together indicate the specifications or standards (such as size) desired by the user. Therefore, if a space s is inserted between numbers, the possibility of extracting unintended products (noise) increases. Specific examples will be given and described below.
[0027] FIG. 6 shows an example where the desired products are products that include "Benri tube" and "13", and the input string n is "Benri tube 13".
[0028] As shown in FIG. 6(a), when a space s is inserted between the same-type characters (here, between numbers and between katakana), the search string k becomes "Ben □ ri □ tube □ 1 □ 3". Then, in addition to the desired product "Benri tube 0784B - 13 * 400mm", unintended products "Benri tube 0784B - 10 * 300mm", "Benri tube 0784B - 30 * 100mm", and "Three Ben tubes 130L" are extracted.
[0029] On the other hand, as shown in Figure 6(b), if no space s is inserted between identical characters, the search string k becomes "Benry□ Pipe□13". In this case, only the desired product "Benry Pipe 0784B-13*400mm" is extracted, and unintended products such as "Benry Pipe 0784B-10*300mm", "Benry Pipe 0784B-30*100mm", and "Three Benry Pipe 130L" are not extracted. In this way, noise can be reduced by not inserting a space s between multiple consecutive identical characters.
[0030] Figure 7 shows an example where the desired product includes "welding" and "long elbow," and the input string n is "welded long elbow."
[0031] As shown in Figure 7(a), when a space s is inserted between identical characters (in this case, between kanji characters, or between katakana characters), the search string k becomes "yō□tsutsu□ro□n□gu□eru□bo". As a result, in addition to the desired product "welding 90L long elbow", an unintended product "melt-bonded long ball 4000 eco type" is also extracted.
[0032] On the other hand, as shown in Figure 7(b), if no space s is inserted between identical characters, the search string k becomes "welding □ long elbow". In this case, only the desired product "welding 90L long elbow" is extracted, and the unintended product "melt-bonded long ball 4000 eco type" is not extracted. In this way, noise can be reduced by not inserting a space s between multiple consecutive identical characters.
[0033] The search string generation unit 13 is configured to delete kana characters if the input string n contains kana characters identical to particles, whether it generates the search string k by inserting a delimiter character after each character of the input string n, or by generating the search string k without inserting a space s between multiple consecutive identical characters of the input string n.
[0034] The search string generation unit 13 can be configured to generate a first search string k1 and a second search string k2 in stages. Here, the first search string k1 represents a string in which, basically, a space s is inserted after each character of the input string n, and no space s is inserted between consecutive identical characters. The second search string k2 represents a string in which a space s is inserted after each character of the input string n. The search string generation unit 13 generates the first search string k1, and the extraction unit 14 uses the first search string k1 to extract products from the database 3. If no products are extracted, the second search string k2 is generated, and the extraction unit 14 uses the second search string k2 to extract products from the database 3. With this configuration, by transitioning from the first search string k1, which has less noise, to the second search string k2, which has more noise but looser extraction conditions, products can be reliably extracted while considering the extraction speed.
[0035] Next, the search process using the search device 1 configured as described above will be explained with reference to Figure 8. Here, we will explain the case in which the search string generation unit 13 generates the first search string k1 and the second search string k2 in stages.
[0036] As shown in Figure 8, when a user enters a search keyword in the search box 22 and taps the search button 23 (S1), the input string acquisition unit 11 acquires the input string n (S2). Next, the extraction attempt unit 12 attempts to extract products from the database 3 using the input string n (S3). If a product is extracted (S4: Yes), the search device 1 displays the extraction result (S10) and terminates the process. On the other hand, if no product is extracted (S4: No), the search string generation unit 13 generates a first search string k1 based on the input string n (S5). Subsequently, the extraction unit 14 extracts products from the database 3 using the first search string k1 (S6).
[0037] If a product is found (S7: Yes), the search device 1 displays the extraction result (S10) and terminates the process. On the other hand, if no product is found (S7: No), the search string generation unit 13 generates a second search string k2 based on the input string n (S8). The extraction unit 14 extracts products from the database 3 using the second search string k2 (S9). The search device 1 displays the extraction result (S10) and terminates the process. It is also possible to display the extraction result after S6 (S10) and terminate the process. Furthermore, it is possible to extract products using only the second search string k2 without using the first search string k1 (omitting the processes in S5-S7).
[0038] Therefore, according to the search device 1 of this embodiment, a space s is inserted for each character of the input string n entered by the user to generate the search string k. As a result, if any of the characters that make up the name of the product are entered in the input string n, the product can be extracted. This makes it easier to find products and search for them easily. In particular, it eliminates the need to re-enter information or correct the input string, even when using voice input.
[0039] Furthermore, this disclosure is not limited to the embodiments described above, and it is possible to implement the invention by appropriately changing the shape and configuration of each part without departing from the spirit of this disclosure. [Explanation of Symbols]
[0040] 1. Search device 2 Information terminals 3 Databases 4 servers 11 Input string acquisition section 12 Extraction trial part 13 Search string generation unit 14 Extraction part 21 monitors 22 Search Box 23 Search button 24 Extraction results N Internet s space n Input string k search string k' Conventional search string
Claims
1. A search device for searching for products registered in a database, An input string acquisition method for obtaining the input string entered by the user, A search string generation means that generates a search string based on the input string, The system includes an extraction means for extracting the product from the database using the search string, The search string generation means is characterized by generating the search string by inserting a delimiter into the input string.
2. The search device according to claim 1, wherein the delimiter is a space.
3. The search device according to claim 1, wherein the search string generation means generates the search string by inserting a delimiter character after each character of the input string.
4. The search device according to claim 1, wherein the search string generation means deletes the kana character if the input string contains the same kana character as a particle.
5. The search device according to claim 1, wherein the search string generation means does not insert the delimiter between a plurality of consecutive identical characters in the input string.
6. The system includes an extraction trial means for attempting to extract the product from the database using the input string, If the product is not extracted by the extraction trial means, the search string generation means generates the search string, and the extraction means extracts the product from the database using the search string, according to claim 1.
7. A search program, which, when executed on a computer, causes the computer to function as the search device described in claim 1.