A search result mining method and device

Abstract

The present invention discloses a method and device for mining search results. The method comprises: obtaining first search result screenshots and second search result screenshots of several search keywords; extracting image features from the first search result screenshots and the second search result screenshots, re-rendering the images according to the extracted image features, and obtaining rendered first search result screenshots and second search result screenshots; clustering the rendered first search result screenshots and the second search result screenshots, and obtaining a category with a higher evaluation value and a category with a lower evaluation value as clustering results; the evaluation value is related to color richness and / or morphological richness; mining search results with lower evaluation values ​​based on the clustering results. The beneficial effect is that a large amount of search results can be automatically mined to obtain search results with lower evaluation values, and the search results of the corresponding search keywords can be subsequently optimized in a targeted manner, which helps to improve user satisfaction.

Description

Technical Field

[0001] The present invention relates to the field of information mining, and in particular to a method and device for mining search results. Background Art

[0002] There is a vast amount of information on the Internet. In order to conveniently obtain information on the Internet and share Internet information resources, people often rely on the help of search engines. However, due to factors such as different search engines using different algorithms and information display methods, different search results will often be displayed when searching for the same search keyword using different search engines. People will also have different evaluations of these search results. In order to achieve continuous optimization and display of massive search results to improve the evaluation value, an automated search result mining method is needed to compare and analyze the search results, thereby identifying search results with lower evaluation values ​​as the basis for subsequent optimization work. Summary of the invention

[0003] In view of the above problems, the present invention is proposed to provide a method and device for mining search results that overcome the above problems or at least partially solve the above problems.

[0004] According to one aspect of the present invention, a method for mining search results is provided, comprising:

[0005] Obtaining screenshots of the first search results and the second search results for several search keywords;

[0006] Extracting image features from the first search result screenshot and the second search result screenshot, and re-rendering the image according to the extracted image features to obtain rendered first search result screenshot and second search result screenshot;

[0007] Clustering the rendered first search result screenshot and the second search result screenshot to obtain a category with a higher evaluation value and a category with a lower evaluation value as clustering results; the evaluation value is related to color richness and / or morphological richness;

[0008] Based on the clustering results, search results with lower evaluation values ​​are mined.

[0009] Optionally, obtaining the first search result screenshots and the second search result screenshots for the plurality of search keywords includes:

[0010] Obtain the search result page of each search keyword in the first search engine, and obtain a first search result screenshot corresponding to each search keyword; and obtain the search result page of each search keyword in the second search engine, and obtain a second search result screenshot corresponding to each search keyword.

[0011] Optionally, the first search result screenshot and the second search result screenshot are obtained by cutting off the head and / or tail of the corresponding search result page after taking screenshots.

[0012] Optionally, the first search engine is a search engine for search results to be mined, and the second search engine is a search engine for reference;

[0013] The method of mining search results with lower evaluation values ​​based on clustering results includes: if a first search result screenshot of a search keyword is in a category with lower evaluation value, and a second search result screenshot is in a category with higher evaluation value, then the search result of the search keyword in the first search engine is used as the search result with lower evaluation value.

[0014] Optionally, extracting image features from the first search result screenshot and the second search result screenshot includes:

[0015] Merge the first search result screenshot and the second search result screenshot to obtain a plurality of merged images;

[0016] Cluster each pixel in the merged image according to color to obtain the color center value of each cluster;

[0017] The re-rendering of the image according to the extracted image features to obtain the rendered first search result screenshot and the second search result screenshot comprises:

[0018] The first search result screenshot and the second search result screenshot are respectively re-rendered according to the color center value.

[0019] Optionally, the merging the first search result screenshot and the second search result screenshot to obtain a plurality of merged images comprises: randomly selecting N search keywords, and merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged image;

[0020] Clustering each pixel in the merged image according to color includes: obtaining N clusters based on a k-means clustering algorithm;

[0021] The re-rendering of each first search result screenshot and second search result screenshot respectively according to the color center value comprises: rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes;

[0022] Wherein, N is a positive integer.

[0023] Optionally, clustering each pixel in the merged image according to color to obtain the color center value of each cluster includes:

[0024] A picture representation of the merged picture is determined based on the RGB value of each pixel, and clustering is performed according to the picture representation.

[0025] According to another aspect of the present invention, a search result mining device is provided, comprising:

[0026] An acquisition unit, adapted to acquire first search result screenshots and second search result screenshots of a plurality of search keywords;

[0027] A rendering unit, adapted to extract image features from the first search result screenshot and the second search result screenshot, and re-render the image according to the extracted image features to obtain rendered first search result screenshot and second search result screenshot;

[0028] An evaluation unit, adapted to cluster the rendered first search result screenshot and the second search result screenshot to obtain a class with a higher evaluation value and a class with a lower evaluation value as clustering results; the evaluation value is related to color richness and / or morphological richness;

[0029] The mining unit is suitable for mining search results with lower evaluation values ​​based on clustering results.

[0030] Optionally, the acquisition unit is suitable for acquiring the search result page of each search keyword in the first search engine and obtaining a first search result screenshot corresponding to each search keyword; and acquiring the search result page of each search keyword in the second search engine and obtaining a second search result screenshot corresponding to each search keyword.

[0031] Optionally, the first search result screenshot and the second search result screenshot are obtained by cutting off the head and / or tail of the corresponding search result page after taking screenshots.

[0032] Optionally, the first search engine is a search engine for search results to be mined, and the second search engine is a search engine for reference;

[0033] The evaluation unit is adapted to treat the search result of the search keyword in the first search engine as the search result with the lower evaluation value if the first search result screenshot of the search keyword is in a category with a lower evaluation value, while the second search result screenshot is in a category with a higher evaluation value.

[0034] Optionally, the rendering unit is suitable for merging the first search result screenshot and the second search result screenshot to obtain several merged pictures; clustering each pixel in the merged picture according to color to obtain the color center value of each cluster; and re-rendering each first search result screenshot and the second search result screenshot according to the color center value.

[0035] Optionally, the rendering unit is suitable for randomly selecting N search keywords, merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged image; obtaining N classes based on the k-means clustering algorithm; and rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes; wherein N is a positive integer.

[0036] Optionally, the rendering unit is adapted to determine a picture representation of the merged picture based on the RGB value of each pixel, and perform clustering according to the picture representation.

[0037] According to another aspect of the present invention, an electronic device is provided, comprising: a processor; and a memory arranged to store computer executable instructions, wherein when the executable instructions are executed, the processor executes any of the above methods.

[0038] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, the one or more programs implement any of the methods described above.

[0039] From the above, it can be seen that the technical solution of the present invention obtains the first search result screenshots and the second search result screenshots of several search keywords; extracts image features from the first search result screenshots and the second search result screenshots, and re-renders the images according to the extracted image features to obtain the rendered first search result screenshots and the second search result screenshots; clusters the rendered first search result screenshots and the second search result screenshots to obtain a category with a higher evaluation value and a category with a lower evaluation value as clustering results; the evaluation value is related to the color richness and / or morphological richness; and mines out search results with lower evaluation values ​​based on the clustering results. The beneficial effect of this technical solution is that it can automatically mine a large number of search results to obtain search results with lower evaluation values, and subsequently the search results of the corresponding search keywords can be optimized in a targeted manner, which helps to improve user satisfaction.

[0040] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0042] Figure 1 A schematic diagram showing a flow chart of a method for mining search results according to an embodiment of the present invention;

[0043] Figure 2 A schematic diagram of the structure of a search result mining device according to an embodiment of the present invention is shown;

[0044] Figure 3 A schematic diagram of the structure of an electronic device according to an embodiment of the present invention is shown;

[0045] Figure 4 A schematic diagram of the structure of a computer-readable storage medium according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0046] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0047] Figure 1 FIG. 2 is a flow chart showing a method for mining search results according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0048] Step S110, obtaining first search result screenshots and second search result screenshots of several search keywords.

[0049] The search keyword query will get different search results under different sorting algorithms and search result display forms. For example, a search result can contain only text or pictures; the search results with the highest ranking can be displayed in various forms or in a single form. The embodiments of the present invention hope to be able to mine based on different search results of the same search keyword. Here, different search results can be obtained based on different search engines, or can be obtained by AB testing in the same search engine, and so on.

[0050] Step S120, extracting image features from the first search result screenshot and the second search result screenshot, and re-rendering the image according to the extracted image features to obtain rendered first search result screenshot and second search result screenshot.

[0051] Since the search results are relatively rich in content and contain a large amount of information, directly processing the search result screenshots is inefficient. Therefore, extracting the image features first and then re-rendering and then processing the obtained images can significantly improve the processing efficiency.

[0052] Step S130, clustering the rendered first search result screenshot and the second search result screenshot to obtain a category with a higher evaluation value and a category with a lower evaluation value as clustering results; the evaluation value is related to color richness and / or morphological richness;

[0053] For example, search results with larger blocks of color values ​​and brighter colors usually have a better user experience, and color features are closely related to morphological features, so the first search result screenshot and the second search result screenshot can be clustered based on the color richness and / or morphological richness as the evaluation value standard. In this way, the factors affecting the user experience can be combined to automatically classify images with different evaluation values ​​according to certain standards, laying the foundation for effectively mining search results with poor user experience and low evaluation values.

[0054] Step S140: mining search results with lower evaluation values ​​based on the clustering results.

[0055] For example, the first search result screenshot of a certain search keyword is in a category with a higher evaluation value, while the second search result screenshot is in a category with a lower evaluation value. In this case, the second search result is the search result with the lower evaluation value, and the corresponding search engine or algorithm can be optimized subsequently.

[0056] It can be seen that Figure 1 The method shown can combine factors that affect user experience and use machine automatic learning methods to mine search results. Images are automatically clustered according to color richness and / or morphological richness, realizing automated analysis and mining in massive search results, and ultimately identifying search results with low evaluation values, laying a solid foundation for targeted optimization and improvement of subsequent search results for corresponding keywords, saving labor costs, greatly improving work efficiency, and helping to improve user satisfaction.

[0057] In an embodiment of the present invention, in the above method, obtaining the first search result screenshots and the second search result screenshots of a plurality of search keywords includes: obtaining the search result pages of each search keyword in the first search engine and obtaining the first search result screenshots corresponding to each search keyword; and obtaining the search result pages of each search keyword in the second search engine and obtaining the second search result screenshots corresponding to each search keyword.

[0058] For example, for the search keyword "言舞", use the first search engine to search, and save a screenshot of the obtained search result page. Then, use the second search engine to search again with "言舞" as the search keyword, and also take a screenshot of the search result page. Correspondingly, by repeating similar operations, finally obtain the screenshots of the search result pages of a plurality of search keywords obtained through different search engines. In this way, the data basis for mining is obtained.

[0059] In an embodiment of the present invention, in the above method, the first search result screenshots and the second search result screenshots are obtained by cutting off the head and / or the tail after taking screenshots of the corresponding search result pages.

[0060] In order to reduce the amount of information analysis and improve work efficiency, some interference factors irrelevant to search result mining can be removed. For example, the head and / or tail parts of the search result page have relatively small differences in display under different keywords, and after removal, the influence of the subtle differences in the head and tail page displays under different search engines can be further reduced. This makes the focus of the search result mining work more concentrated, and the efficiency of the mining work can also be further improved after removing these interference items such as the head and / or tail.

[0061] In an embodiment of the present invention, in the above method, the first search engine is the search engine for which the search results are to be mined, and the second search engine is the reference search engine; mining out the search results with relatively low evaluation values based on the clustering results includes: if the first search result screenshot of a search keyword is in a category with relatively low evaluation values, while the second search result screenshot is in a category with relatively high evaluation values, then take the search result of this search keyword in the first search engine as the search result with relatively low evaluation values.

[0062] Searching for the same keyword will yield different search results. The first search result can be obtained from the search engine to be mined, and the second search result is obtained from the reference search engine. For example, when searching for the same keyword "Yanwu" on different search engines, two types of search results can be obtained, one is the search result under the search engine to be mined, and the other is the search result under the reference search engine. Then, subsequent processing is carried out using the color richness and / or morphological richness as the evaluation value criteria. If the search result of "Yanwu" under the search engine to be mined is classified into the category with a lower evaluation value, while the evaluation value of the search result of "Yanwu" under the reference search engine is classified into the category with a higher evaluation value, then it shows that for the search keyword "Yanwu", the evaluation value of the search result of the second search engine is higher than that of the first search engine. The first search engine can optimize algorithms and display forms for "Yanwu" or similar search keywords.

[0063] In an embodiment of the present invention, in the above method, extracting image features from the first search result screenshot and the second search result screenshot includes: performing a merging process on the first search result screenshot and the second search result screenshot to obtain several merged images; clustering each pixel point in the merged images according to color to obtain the color center values of each cluster; re-rendering the images according to the extracted image features to obtain the rendered first search result screenshot and the second search result screenshot includes: re-rendering each first search result screenshot and the second search result screenshot respectively according to the color center values.

[0064] To obtain better results, when extracting pixel point features, several screenshots of the search results of the search engine to be mined and the reference search engine screenshots can be merged first, and then pixel point features can be extracted from the merged images. Clustering the merged images according to color features to obtain the color median values of each cluster. For example, for the color yellow in the merged image, there are 5 types in total, namely A1, A2, A3, A4, and A5. Among them, the yellow features of A1 and A3 are relatively close, while the yellow of A2, A4, and A5 are relatively close. At this time, after clustering based on a certain algorithm, two types of yellow will be obtained. A1 and A3 are in one cluster, and the color median value A13 is obtained; A2, A4, and A5 are in one cluster, and the color median value A245 is obtained. When rendering the original image, A1 and A3 in the original image are re-rendered as A13, and A2, A4, and A5 are re-rendered as A245. This overcomes the problem that when clustering the color features extracted from a single image, they will still be considered as images with relatively large distinguishing features and thus cannot be well clustered.

[0065] In one embodiment of the present invention, in the above method, the first search result screenshot and the second search result screenshot are merged to obtain a number of merged pictures, including: randomly selecting N search keywords, and merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged picture; clustering each pixel in the merged picture according to color includes: obtaining N classes based on the k-means clustering algorithm; re-rendering each first search result screenshot and second search result screenshot respectively according to the color center value includes: rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes; wherein N is a positive integer.

[0066] For example, there are 32 keywords used for search result mining, and 64 search results are obtained based on different search engines. These 64 images are merged, and then the color features are clustered. The clustering algorithm can be the kmeans mean clustering algorithm. Then, the 32 first search result screenshots and 32 second search result screenshots are re-rendered using the intermediate values ​​of the color features obtained after clustering.

[0067] In one embodiment of the present invention, in the above method, clustering each pixel in the merged image according to color to obtain the color center value of each cluster includes: determining the image representation of the merged image based on the RGB value of each pixel, and clustering according to the image representation.

[0068] The color information of the image is represented by the RGB value of the pixel information, and then clustered based on the feature information of all the pixels. This converts the user's experience of different color features into a language that the computer can understand, which is convenient for effective execution and improvement, and this processing method is relatively efficient.

[0069] Figure 2 FIG. 2 is a schematic diagram showing a structure of a search result mining device according to an embodiment of the present invention. Figure 2 As shown, the search result mining device 200 includes:

[0070] An acquisition unit 210 is adapted to acquire first search result screenshots and second search result screenshots of a plurality of search keywords;

[0071] The search keyword query will get different search results under different sorting algorithms and search result display forms. For example, a search result can contain only text or pictures; the search results with the highest ranking can be displayed in various forms or in a single form. The embodiments of the present invention hope to be able to mine based on different search results of the same search keyword. Here, different search results can be obtained based on different search engines, or can be obtained by AB testing in the same search engine, and so on.

[0072] The rendering unit 220 is adapted to extract image features from the first search result screenshot and the second search result screenshot, and re-render the image according to the extracted image features to obtain the rendered first search result screenshot and the second search result screenshot;

[0073] Since the search results are relatively rich in content and contain a large amount of information, directly processing the search result screenshots is inefficient. Therefore, extracting the image features first and then re-rendering and then processing the obtained images can significantly improve the processing efficiency.

[0074] An evaluation unit 230 is adapted to cluster the rendered first search result screenshot and the second search result screenshot to obtain a class with a higher evaluation value and a class with a lower evaluation value as clustering results; the evaluation value is related to color richness and / or morphological richness;

[0075] For example, search results with larger blocks of color values ​​and brighter colors usually have a better user experience, and color features are closely related to morphological features, so the first search result screenshot and the second search result screenshot can be clustered based on the color richness and / or morphological richness as the evaluation value standard. In this way, the factors affecting the user experience can be combined to automatically classify images with different evaluation values ​​according to certain standards, laying the foundation for effectively mining search results with poor user experience and low evaluation values.

[0076] The mining unit 240 is adapted to mine search results with lower evaluation values ​​based on the clustering results.

[0077] For example, the first search result screenshot of a certain search keyword is in a category with a higher evaluation value, while the second search result screenshot is in a category with a lower evaluation value. In this case, the second search result is the search result with the lower evaluation value, and the corresponding search engine or algorithm can be optimized subsequently.

[0078] It can be seen that Figure 2 The device shown can combine factors that affect user experience and use machine automatic learning methods to mine search results. It automatically clusters images according to color richness and / or morphological richness, realizes automated analysis and mining in massive search results, and ultimately identifies search results with low evaluation values, laying a solid foundation for subsequent targeted optimization and improvement in massive search results, saving labor costs, greatly improving work efficiency, and helping to improve user satisfaction.

[0079] In one embodiment of the present invention, in the above device, the acquisition unit 210 is adapted to acquire the search result pages of each search keyword in the first search engine and obtain the first search result screenshots corresponding to each search keyword; and acquire the search result pages of each search keyword in the second search engine and obtain the second search result screenshots corresponding to each search keyword.

[0080] For example, for "Yanwu" as the search keyword, use the first search engine to search, and save the screenshot of the obtained search result page. Then, use the second search engine to search again with "Yanwu" as the search keyword, and also take a screenshot of the search result page. By repeating similar operations accordingly, finally obtain the screenshots of the search result pages obtained by different search engines for several search keywords. In this way, the data basis for mining is obtained.

[0081] In one embodiment of the present invention, in the above device, the first search result screenshot and the second search result screenshot are obtained by cutting off the head and / or the tail after taking screenshots of the corresponding search result pages.

[0082] In order to reduce the amount of information analysis and improve work efficiency, some interference factors irrelevant to search result mining can be removed. For example, the head and / or tail parts of the search result page have relatively small differences in display under different keywords, and after removal, the influence of the subtle differences in the head and tail page displays under different search engines can be further reduced. This makes the focus of the search result mining work more concentrated, and the efficiency of the mining work can also be further improved after removing these interference items such as the head and / or tail.

[0083] In one embodiment of the present invention, in the above device, the first search engine is the search engine for which the search results are to be mined, and the second search engine is the search engine for reference; the evaluation unit 230 is adapted to, if the first search result screenshot of a search keyword is in a category with a lower evaluation value, while the second search result screenshot is in a category with a higher evaluation value, then regard the search result of the search keyword in the first search engine as the search result with a lower evaluation value.

[0084] Searching for the same keyword will yield different search results. The first search result can be obtained through the search engine to be mined, and the second search result is obtained through the reference search engine. For example, when searching for the same keyword "言舞" in different search engines, two types of search results can be obtained, one is the search result under the search engine to be mined, and the other is the search result under the reference search engine. Then, subsequent processing is carried out according to the color richness and / or morphological richness as the evaluation value criteria. If the search result of "言舞" under the search engine to be mined is classified into a category with a lower evaluation value, while the evaluation value of the search result of "言舞" under the reference search engine is classified into a category with a higher evaluation value, then it shows that for the search keyword "言舞", the evaluation value of the search result of the second search engine is higher than that of the first search engine. The first search engine can optimize algorithms and display forms for "言舞" or similar search keywords.

[0085] In an embodiment of the present invention, in the above device, the rendering unit 220 is adapted to perform a merging process on the screenshots of the first search result and the screenshots of the second search result to obtain several merged pictures; cluster the pixel points in the merged pictures according to colors to obtain the color center values of each cluster; and is adapted to re-render each screenshot of the first search result and the screenshots of the second search result according to the color center values.

[0086] To obtain better results, when extracting pixel point features, several screenshots of the search results of the search engine to be mined and the screenshots of the search results of the reference search engine can be first merged, and then the pixel point features of the merged pictures are extracted. Cluster the merged pictures according to color features to obtain the color median values of each cluster. For example, for the color yellow in the merged picture, there are a total of 5 types, namely A1, A2, A3, A4, and A5. Among them, the yellow features of A1 and A3 are relatively close, while the three yellows of A2, A4, and A5 are relatively close. At this time, after clustering based on a certain algorithm, two types of yellows will be obtained. A1 and A3 are in one cluster, and the color median value A13 is obtained; A2, A4, and A5 are in one cluster, and the color median value A245 is obtained. When rendering the original picture, A1 and A3 in the original picture are re-rendered as A13, and A2, A4, and A5 are re-rendered as A245. This overcomes the problem that when clustering the color features extracted from a single picture, the pictures are still considered to have relatively large distinguishing features and thus cannot be well clustered.

[0087] In one embodiment of the present invention, in the above-mentioned device, the rendering unit 220 is suitable for randomly selecting N search keywords, merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged image; obtaining N classes based on the k-means clustering algorithm; rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes; wherein N is a positive integer.

[0088] For example, there are 32 keywords used for search result mining, and 64 search results are obtained based on different search engines. These 64 images are merged, and then the color features are clustered. The clustering algorithm can be the kmeans mean clustering algorithm. Then, the 32 first search result screenshots and 32 second search result screenshots are re-rendered using the intermediate values ​​of the color features obtained after clustering.

[0089] In one embodiment of the present invention, in the above-mentioned device, the rendering unit 220 is adapted to determine a picture representation of the merged picture based on the RGB value of each pixel, and perform clustering according to the picture representation.

[0090] The color information of the image is represented by the RGB value of the pixel information, and then clustered based on the feature information of all the pixels. This converts the user's experience of different color features into a language that the computer can understand, which is convenient for effective execution and improvement, and this processing method is relatively efficient.

[0091] In summary, the technical solution of the present invention obtains the first search result screenshot and the second search result screenshot of several search keywords; extracts picture features from the first search result screenshot and the second search result screenshot, re-renders the picture according to the extracted picture features, and obtains the rendered first search result screenshot and the second search result screenshot; clusters the rendered first search result screenshot and the second search result screenshot, and obtains a class with a higher evaluation value and a class with a lower evaluation value as clustering results; the evaluation value is related to the color richness and / or the morphology richness; and mines the search results with a lower evaluation value based on the clustering results. The beneficial effect of this technical solution is that the search results can be mined by combining the factors affecting the user experience with the method of machine automatic learning. The pictures are automatically clustered according to the color richness and / or the morphology richness as the standard, and the automatic analysis and mining work in the massive search results is realized, and the search results with a lower evaluation value are finally identified. It lays a good foundation for the targeted optimization and improvement of the search results of the corresponding keywords in the future, saves labor costs, greatly improves work efficiency, and helps to improve user satisfaction.

[0092] It should be noted that:

[0093] The algorithm and display provided herein are not inherently related to any particular computer, virtual device or other equipment. Various general-purpose devices can also be used together with the teachings based on this. According to the above description, it is obvious to construct the structure required for this type of device. In addition, the present invention is not directed to any specific programming language either. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description of the above specific language is for the purpose of disclosing the best mode of the present invention.

[0094] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0095] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, the various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: that the claimed invention requires more features than the features explicitly recited in each claim. More specifically, as reflected in the claims below, inventive aspects lie in less than all the features of the individual embodiments disclosed above. Therefore, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, with each claim itself serving as a separate embodiment of the present invention.

[0096] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0097] In addition, those skilled in the art will appreciate that, although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present invention and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.

[0098] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in the mining device for the search results according to an embodiment of the present invention. The present invention may also be implemented as a device or device program (e.g., a computer program and a computer program product) for executing part or all of the methods described herein. Such a program for implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0099] For example, Figure 3 A schematic diagram of the structure of an electronic device according to an embodiment of the present invention is shown. The electronic device 300 includes a processor 310 and a memory 320 arranged to store computer executable instructions (computer readable program code). The memory 320 can be an electronic memory such as a flash memory, an EEPROM (electrically erasable programmable read-only memory), an EPROM, a hard disk or a ROM. The memory 320 has a storage space 330 for storing a computer readable program code 331 for executing any method step in the above method. For example, the storage space 330 for storing computer readable program code may include individual computer readable program codes 331 for implementing various steps in the above method respectively. The computer readable program code 331 can be read from or written into one or more computer program products. These computer program products include program code carriers such as a hard disk, a compact disk (CD), a memory card or a floppy disk. Such a computer program product is typically, for example Figure 4 The computer-readable storage medium. Figure 4A schematic diagram of the structure of a computer-readable storage medium according to an embodiment of the present invention is shown. The computer-readable storage medium 400 stores a computer-readable program code 331 for executing the method steps according to the present invention, which can be read by the processor 310 of the electronic device 300. When the computer-readable program code 331 is executed by the electronic device 300, the electronic device 300 executes the various steps in the method described above. Specifically, the computer-readable program code 331 stored in the computer-readable storage medium can execute the method shown in any of the above embodiments. The computer-readable program code 331 can be compressed in an appropriate form.

[0100] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.

[0101] The embodiment of the present invention discloses A1, a method for mining search results, comprising:

[0102] Obtaining screenshots of the first search results and the second search results for several search keywords;

[0103] Extracting image features from the first search result screenshot and the second search result screenshot, and re-rendering the image according to the extracted image features to obtain rendered first search result screenshot and second search result screenshot;

[0104] Clustering the rendered first search result screenshot and the second search result screenshot to obtain a category with a higher evaluation value and a category with a lower evaluation value as clustering results; the evaluation value is related to color richness and / or morphological richness;

[0105] Based on the clustering results, search results with lower evaluation values ​​are mined.

[0106] A2. The method as described in A1, wherein obtaining the first search result screenshots and the second search result screenshots for the plurality of search keywords comprises:

[0107] Obtain the search result page of each search keyword in the first search engine, and obtain a first search result screenshot corresponding to each search keyword; and obtain the search result page of each search keyword in the second search engine, and obtain a second search result screenshot corresponding to each search keyword.

[0108] A3. A method as described in A2, wherein the first search result screenshot and the second search result screenshot are obtained by cutting off the head and / or tail of the corresponding search result page after taking screenshots.

[0109] A4. The method as described in A2, wherein the first search engine is a search engine for search results to be mined, and the second search engine is a search engine for reference;

[0110] The method of mining search results with lower evaluation values ​​based on clustering results includes: if a first search result screenshot of a search keyword is in a category with lower evaluation value, and a second search result screenshot is in a category with higher evaluation value, then the search result of the search keyword in the first search engine is used as the search result with lower evaluation value.

[0111] A5. The method as described in A1, wherein extracting image features from the first search result screenshot and the second search result screenshot comprises:

[0112] Merge the first search result screenshot and the second search result screenshot to obtain a plurality of merged images;

[0113] Cluster each pixel in the merged image according to color to obtain the color center value of each cluster;

[0114] The re-rendering of the image according to the extracted image features to obtain the rendered first search result screenshot and the second search result screenshot comprises:

[0115] The first search result screenshot and the second search result screenshot are respectively re-rendered according to the color center value.

[0116] A6. The method as described in A5, wherein the merging the first search result screenshot and the second search result screenshot to obtain a plurality of merged images comprises: randomly selecting N search keywords, merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged image;

[0117] Clustering each pixel in the merged image according to color includes: obtaining N clusters based on a k-means clustering algorithm;

[0118] The re-rendering of each first search result screenshot and second search result screenshot respectively according to the color center value comprises: rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes;

[0119] Wherein, N is a positive integer.

[0120] A7. The method as described in A5, wherein clustering the pixels in the merged image according to color to obtain the color center values ​​of each cluster includes:

[0121] A picture representation of the merged picture is determined based on the RGB value of each pixel, and clustering is performed according to the picture representation.

[0122] The embodiment of the present invention discloses B8, a search result mining device, comprising:

[0123] An acquisition unit, adapted to acquire first search result screenshots and second search result screenshots of a plurality of search keywords;

[0124] A rendering unit, adapted to extract image features from the first search result screenshot and the second search result screenshot, and re-render the image according to the extracted image features to obtain rendered first search result screenshot and second search result screenshot;

[0125] An evaluation unit, adapted to cluster the rendered first search result screenshot and the second search result screenshot to obtain a class with a higher evaluation value and a class with a lower evaluation value as clustering results; the evaluation value is related to color richness and / or morphological richness;

[0126] The mining unit is suitable for mining search results with lower evaluation values ​​based on clustering results.

[0127] B9. The device as described in B8, wherein:

[0128] The acquisition unit is adapted to acquire the search result page of each search keyword in the first search engine and obtain a first search result screenshot corresponding to each search keyword; and acquire the search result page of each search keyword in the second search engine and obtain a second search result screenshot corresponding to each search keyword.

[0129] B10. The device as described in B9, wherein the first search result screenshot and the second search result screenshot are obtained by cutting off the head and / or tail of the corresponding search result page after taking screenshots.

[0130] B11. The device as described in B10, wherein the first search engine is a search engine for search results to be mined, and the second search engine is a search engine for reference;

[0131] The evaluation unit is adapted to treat the search result of the search keyword in the first search engine as the search result with the lower evaluation value if the first search result screenshot of the search keyword is in a category with a lower evaluation value, while the second search result screenshot is in a category with a higher evaluation value.

[0132] B12. The device as described in B8, wherein:

[0133] The rendering unit is suitable for merging the first search result screenshot and the second search result screenshot to obtain a plurality of merged images; clustering each pixel in the merged image according to color to obtain the color center value of each cluster; and re-rendering each first search result screenshot and the second search result screenshot respectively according to the color center value.

[0134] B13. The device as described in B12, wherein:

[0135] The rendering unit is suitable for randomly selecting N search keywords, merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged image; obtaining N classes based on the k-means clustering algorithm; and rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes; wherein N is a positive integer.

[0136] B14. The device as described in B12, wherein:

[0137] The rendering unit is adapted to determine a picture representation of the merged picture based on the RGB value of each pixel, and perform clustering according to the picture representation.

[0138] An embodiment of the present invention discloses C15, an electronic device, wherein the electronic device comprises: a processor; and a memory arranged to store computer executable instructions, wherein the executable instructions, when executed, cause the processor to execute a method as described in any one of A1-A7.

[0139] An embodiment of the present invention discloses D16, a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and when the one or more programs are executed by a processor, implements the method as described in any one of A1-A7.

Claims

1. A method for mining search results, comprising: Obtaining screenshots of the first search results and the second search results for several search keywords; Extracting image features from the first search result screenshot and the second search result screenshot, and re-rendering the image according to the extracted image features to obtain rendered first search result screenshot and second search result screenshot; Clustering the rendered first search result screenshot and the second search result screenshot to obtain a category with a higher evaluation value and a category with a lower evaluation value as clustering results; The evaluation value is related to color richness and / or morphological richness; Based on the clustering results, search results with lower evaluation values ​​are mined.

2. The method of claim 1, wherein: The obtaining of first search result screenshots and second search result screenshots of a plurality of search keywords comprises: Obtain the search result page of each search keyword in the first search engine, and obtain the first search result screenshot corresponding to each search keyword; and obtain the search result page of each search keyword in the second search engine, and obtain the second search result screenshot corresponding to each search keyword.

3. The method of claim 2, wherein: The first search result screenshot and the second search result screenshot are obtained by cutting off the head and / or tail of the corresponding search result page after taking screenshots.

4. The method of claim 2, wherein: The first search engine is a search engine for search results to be mined, and the second search engine is a search engine used for reference; The method of mining search results with lower evaluation values ​​based on clustering results includes: if a first search result screenshot of a search keyword is in a category with lower evaluation value, and a second search result screenshot is in a category with higher evaluation value, then the search result of the search keyword in the first search engine is used as the search result with lower evaluation value.

5. The method of claim 1, wherein: The extracting of image features from the first search result screenshot and the second search result screenshot comprises: Merge the first search result screenshot and the second search result screenshot to obtain a plurality of merged images; Cluster each pixel in the merged image according to color to obtain the color center value of each cluster; The re-rendering of the image according to the extracted image features to obtain the rendered first search result screenshot and the second search result screenshot comprises: The first search result screenshot and the second search result screenshot are respectively re-rendered according to the color center value.

6. The method of claim 5, wherein: The merging of the first search result screenshot and the second search result screenshot to obtain a plurality of merged images includes: randomly selecting N search keywords, and merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged image; Clustering each pixel in the merged image according to color includes: obtaining N clusters based on a k-means clustering algorithm; The re-rendering of each first search result screenshot and second search result screenshot respectively according to the color center value comprises: rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes; Wherein, N is a positive integer.

7. The method of claim 5, wherein: The pixel points in the merged image are clustered according to color to obtain the color center values ​​of each cluster, including: A picture representation of the merged picture is determined based on the RGB value of each pixel, and clustering is performed according to the picture representation.

8. A search result mining device, comprising: An acquisition unit, adapted to acquire first search result screenshots and second search result screenshots of a plurality of search keywords; A rendering unit, adapted to extract image features from the first search result screenshot and the second search result screenshot, and re-render the image according to the extracted image features to obtain rendered first search result screenshot and second search result screenshot; An evaluation unit, adapted to cluster the rendered first search result screenshot and the second search result screenshot to obtain a category with a higher evaluation value and a category with a lower evaluation value as clustering results; The evaluation value is related to color richness and / or morphological richness; The mining unit is suitable for mining search results with lower evaluation values ​​based on clustering results.

9. The device of claim 8, wherein: The acquisition unit is adapted to acquire the search result page of each search keyword in the first search engine and obtain a first search result screenshot corresponding to each search keyword; and acquire the search result page of each search keyword in the second search engine and obtain a second search result screenshot corresponding to each search keyword.

10. The device of claim 9, wherein: The first search result screenshot and the second search result screenshot are obtained by cutting off the head and / or tail of the corresponding search result page after taking screenshots.

11. The device of claim 10, wherein: The first search engine is a search engine for search results to be mined, and the second search engine is a search engine used for reference; The evaluation unit is adapted to treat the search result of the search keyword in the first search engine as the search result with the lower evaluation value if the first search result screenshot of the search keyword is in a category with a lower evaluation value, while the second search result screenshot is in a category with a higher evaluation value.

12. The device of claim 8, wherein: The rendering unit is suitable for merging the first search result screenshot and the second search result screenshot to obtain a plurality of merged images; clustering each pixel in the merged image according to color to obtain the color center value of each cluster; and re-rendering each first search result screenshot and the second search result screenshot respectively according to the color center value.

13. The device of claim 12, wherein: The rendering unit is suitable for randomly selecting N search keywords, merging the corresponding N first search result screenshots and N second search result screenshots to obtain a merged image; obtaining N classes based on the k-means clustering algorithm; and rendering the N first search result screenshots and the N second search result screenshots respectively according to the color center values ​​of the N classes; wherein N is a positive integer.

14. The device of claim 12, wherein: The rendering unit is adapted to determine a picture representation of the merged picture based on the RGB value of each pixel, and perform clustering according to the picture representation.

15. An electronic device, wherein: The electronic device comprises: a processor; and a memory arranged to store computer executable instructions, wherein when the executable instructions are executed, the processor executes the method according to any one of claims 1 to 7.

16. A computer-readable storage medium, wherein: The computer-readable storage medium stores one or more programs, which, when executed by a processor, implement the method according to any one of claims 1 to 7.

Images