A collaborative filtering recommendation algorithm based on improved user similarity

Abstract

The present application relates to the technical field of collaborative filtering recommendation algorithm, in particular to a collaborative filtering recommendation algorithm based on improved user similarity, comprising the following steps: obtaining a task group target, judging the type of the task group target, and analyzing the corresponding use scene of the type of the task group target; classifying the task group target according to the fault setting result of the task group target, and referring to the classification of the task target group to regulate the pitch, tone and loudness of the voice assistant program related in the hardware; matching the regulation result of the pitch, tone and loudness of the voice assistant program related in the hardware to the corresponding use scene; the present application can periodically and autonomously complete the alternation of recommended items, and can better fit the user's usage habits. At the same time, the top control logic program in the present application can assist the present application, and when in use, a variety of recommendation schemes are provided for different people to use, so that the compatibility of the present application is further improved.

Description

Technical Field

[0001] This invention relates to the field of collaborative filtering recommendation algorithm technology, and specifically to a collaborative filtering recommendation algorithm based on improved user similarity. Background Technology

[0002] Intelligent voice technology is a manifestation of artificial intelligence and is now widely used in shopping mall entrances, hospitals, and business service venues. It replaces human intervention by providing basic functions such as answering questions, providing information, and guiding visitors to their destinations.

[0003] However, while existing intelligent voice robots can complete the service tasks set in the program, they cannot proactively provide service assistance based on user needs. This requires users to search or ask questions through dialogue, which reduces the efficiency and intelligence of the intelligent voice robot to some extent. This also shows that there is still room for improvement in the related technologies of intelligent voice robots. Summary of the Invention

[0004] Technical problems to be solved

[0005] To address the aforementioned shortcomings of existing technologies, this invention provides a collaborative filtering recommendation algorithm based on improved user similarity. This algorithm solves the problem that while existing intelligent voice robots can complete the service tasks programmed into their programs, they cannot proactively provide service assistance based on user needs. Consequently, users are required to search or ask questions through dialogue, which reduces the efficiency and intelligence of the intelligent voice robot to some extent.

[0006] Technical solution

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] Firstly, a collaborative filtering recommendation algorithm based on improved user similarity includes the following steps:

[0009] Step 1: Obtain the target of the task group, determine the type of the target of the task group, and analyze the corresponding use cases of the target type of the task group;

[0010] Step 2: Classify the task group targets based on the task group target segmentation setting results, and adjust the relevant pitch, timbre and loudness of the voice assistant program in the hardware with reference to the task target group classification;

[0011] Step 3: Match the adjustment results of the voice assistant program in the hardware with the corresponding usage scenario;

[0012] Step 4: Periodically collect user demand data, and simultaneously evaluate user attributes, corresponding usage scenarios and the matching degree of related voice assistant adjustments based on the collected data;

[0013] Step 5: Set the matching degree assessment qualification index, and capture the user's similar operation behavior characteristics after confirming that the matching degree assessment meets the qualification index.

[0014] Step 6: Generate user demand prediction items based on the captured similar user operation behavior characteristics, send the user demand prediction items to the initial recommendation items and set them to the top;

[0015] Step 7: Users can set the detection cycle to statistically calculate the frequency of use of the top user demand prediction items and set the usage indicators.

[0016] Step 8: Filter the predicted user demand items according to the set usage indicators, and record the filtered items as the preferred recommended user demand items;

[0017] Step 9: Obtain the most frequently used recommended item under the top user demand prediction item, set the most frequently used recommended item as the preferred recommended item, and replace the user demand prediction item with the preferred recommended item for top placement. Set the top placement control logic program.

[0018] Furthermore, in Step 1, the number of usage scenario items corresponding to the target type of the task group is ≥1, where the number of usage scenario items is a natural number.

[0019] Furthermore, Step 1 includes sub-steps, comprising the following steps:

[0020] Step 11: Collect the target requirements of the task group and collect the target attributes of the task group;

[0021] Step 12: Set task group target breaks based on the collected task group target attributes;

[0022] Step 11 and Step 12 are specifically used for determining the target type of the task group in Step 1 and the corresponding application scenario.

[0023] Furthermore, the reference for adjusting the tone, timbre, and loudness of the voice assistant program in the hardware in Step 2 also includes the usage scenarios corresponding to the target types of the task groups analyzed in Step 1.

[0024] Furthermore, in Step 4, the period for collecting user demand data is set to ≤ seven days.

[0025] Furthermore, in Step 4, a user demand initialization recommendation item is set. The user demand initialization recommendation item is initially set based on the usage scenario corresponding to the task group target type analyzed in Step 1 and the task group target segmentation results set in Step 12.

[0026] Furthermore, if the matching degree assessment does not meet the assessment qualification standard, Step 5 will execute Step 11 again and send the processing result of Step 11 to Step 12. Step 12 will refer to the content transmitted in Step 11 to re-call back the target users in each category of the task group target.

[0027] Furthermore, in Step 7, the statistical logic is the ratio between the usage frequency of each item in the user demand forecast item and the number of users; the calculation logic is the ratio between the usage frequency of all items in the user demand forecast item and the number of days the user demand forecast item is deployed.

[0028] Secondly, the top-level control logic program in Step 9 includes:

[0029] The main control module is the main control terminal of the top control logic program, used to control the program's operation and the reception and transmission of instructions;

[0030] The recording module is used to record the click-through rate of users using recommended items;

[0031] The selection module is used to obtain the click-through rate results of user usage of recommended items in the record module, select suitable recommended items and the corresponding suitable recommended item replacement target of the current preferred recommended item;

[0032] The iteration module is used to replace the selection target of suitable recommendations with the preferred recommendations.

[0033] Furthermore, an extended application module is provided between the recording module and the iteration module, including:

[0034] The identification module is used to identify the target attributes of the task group;

[0035] The settings module is used to set the preferred recommendations for the target attributes of each type of task group;

[0036] The jump module is used to jump to the preferred recommendation option corresponding to the target attribute of the task group in the setting module after the identification module identifies the target attribute of the task group.

[0037] Beneficial effects

[0038] Compared with known public technologies, the technical solution provided by this invention has the following beneficial effects:

[0039] 1. This invention can adjust the intelligent voice robot according to the usage scenario and the target audience, making the intelligent voice robot more in line with people's usage needs.

[0040] 2. Through practical use, this invention enables intelligent voice robots equipped with this invention to possess certain learning and judgment abilities, thereby better serving users.

[0041] 3. When in use, the present invention can periodically and autonomously update the recommended items, thereby ensuring that the intelligent voice robot equipped with the present invention keeps up with the times and is more in line with the user's usage habits. At the same time, the top control logic program in the present invention can assist the present invention, providing multiple recommendation schemes for different groups of people to use, thus further improving the compatibility of the present invention. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of this invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0043] Figure 1 This is a schematic diagram of the structure of a collaborative filtering recommendation algorithm based on improved user similarity;

[0044] Figure 2 This is a schematic diagram of the top-level control logic program;

[0045] The labels in the diagram represent: 1. Main control module; 2. Recording module; 21. Identification module; 22. Setting module; 23. Jump module; 3. Selection module; 4. Iteration module. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0047] The present invention will be further described below with reference to embodiments.

[0048] Example 1

[0049] This embodiment presents a collaborative filtering recommendation algorithm based on improved user similarity, such as... Figure 1 As shown, it includes the following steps:

[0050] Step 1: Obtain the target of the task group, determine the type of the target of the task group, and analyze the corresponding use cases of the target type of the task group;

[0051] Step 2: Classify the task group targets based on the task group target segmentation setting results, and adjust the relevant pitch, timbre and loudness of the voice assistant program in the hardware with reference to the task target group classification;

[0052] Step 3: Match the adjustment results of the voice assistant program in the hardware with the corresponding usage scenario;

[0053] Step 4: Periodically collect user demand data, and simultaneously evaluate user attributes, corresponding usage scenarios and the matching degree of related voice assistant adjustments based on the collected data;

[0054] Step 5: Set the matching degree assessment qualification index, and capture the user's similar operation behavior characteristics after confirming that the matching degree assessment meets the qualification index.

[0055] Step 6: Generate user demand prediction items based on the captured similar user operation behavior characteristics, send the user demand prediction items to the initial recommendation items and set them to the top;

[0056] Step 7: Users can set the detection cycle to statistically calculate the frequency of use of the top user demand prediction items and set the usage indicators.

[0057] Step 8: Filter the predicted user demand items according to the set usage indicators, and record the filtered items as the preferred recommended user demand items;

[0058] Step 9: Obtain the most frequently used recommended item under the top user demand prediction item, set the most frequently used recommended item as the preferred recommended item, and replace the user demand prediction item with the preferred recommended item for top placement. Set the top placement control logic program.

[0059] Example 2

[0060] At the implementation level, based on Example 1, this example refers to... Figure 1 The following provides a further detailed explanation of the collaborative filtering recommendation algorithm for improving user similarity in Example 1, as shown below. Figure 1 As shown, in Step 1, the number of usage scenario items corresponding to the task group target type analysis is ≥ 1, where the number of usage scenario items is a natural number.

[0061] like Figure 1 As shown, Step 1 includes sub-steps, including the following steps:

[0062] Step 11: Collect the target requirements of the task group and collect the target attributes of the task group;

[0063] Step 12: Set task group target breaks based on the collected task group target attributes;

[0064] Step 11 and Step 12 are specifically used for determining the target type of the task group in Step 1 and the corresponding application scenario.

[0065] like Figure 1 As shown, the reference for adjusting the tone, timbre and loudness of the voice assistant program in the hardware in Step 2 also includes the usage scenarios corresponding to the task group target types analyzed in Step 1.

[0066] like Figure 1 As shown, in Step 4, the period for collecting user demand data is set to ≤ seven days.

[0067] like Figure 1 As shown, Step 4 includes a user demand initialization recommendation item. This item is initially set based on the task group target type and corresponding usage scenario analyzed in Step 1, as well as the task group target segmentation results set in Step 12.

[0068] like Figure 1 As shown, if the matching degree assessment does not meet the assessment qualification standard, Step 5 will execute Step 11 again and send the processing result of Step 11 to Step 12. Step 12 will refer to the content transmitted in Step 11 to re-call back the target users in each category of the task group target.

[0069] like Figure 1 As shown, the statistical logic in Step 7 is the ratio between the usage frequency of each item in the user demand forecast item and the number of users; the calculation logic is the ratio between the usage frequency of all items in the user demand forecast item and the number of days the user demand forecast item is deployed.

[0070] Example 3

[0071] At the implementation level, based on Example 1, this example refers to... Figure 1 The following provides a further detailed explanation of the collaborative filtering recommendation algorithm for improving user similarity in Example 1, as shown below. Figure 2 As shown, the top-level control logic program in Step 9 includes:

[0072] Main control module 1 is the main control terminal of the top control logic program, used to control program operation and the reception and transmission of instructions;

[0073] Recording module 2 is used to record the click-through rate of users using recommended items;

[0074] Select module 3 to obtain the click-through rate results of user use of recommended items in the record module 2, select suitable recommended items and the corresponding suitable recommended item replacement target of the current preferred recommended item;

[0075] The iteration module 4 is used to replace the selection targets of suitable recommendations and preferred recommendations.

[0076] In this embodiment, the main control module 1 controls the recording module 2 to start recording the click rate of the user's use of the recommended items. The selection module 3 obtains the click rate results of the user's use of the recommended items in the recording module 2, selects the appropriate recommended items and the corresponding appropriate recommended item replacement target of the current preferred recommended items, and finally the replacement module 4 replaces the selected target of the appropriate recommended items with the preferred recommended items, and the replacement of the preferred recommended items is completed.

[0077] like Figure 2 As shown, an extended application module is provided between the recording module 2 and the iteration module 4, including:

[0078] Identification module 21 is used to identify the target attributes of the task group;

[0079] Setting module 22 is used to set the preferred recommendation items corresponding to the target attributes of each type of task group;

[0080] The jump module 23 is used to jump to the preferred recommendation item corresponding to the target attribute of the task group in the setting module 22 after the identification module 21 identifies the target attribute of the task group.

[0081] When in use, the recognition module 21 identifies the target attributes of the task group, and the jump module 23 assists the jump module 23 in jumping to the corresponding target attribute of the task group in the setting module 22 after the recognition module 21 identifies the target attributes of the task group. In this way, the intelligent voice robot can provide more suitable target recommendations for people with different attributes.

[0082] In summary, this invention can adjust the intelligent voice robot according to the usage scenario and the target audience, making the intelligent voice robot more in line with people's usage needs.

[0083] Furthermore, through practical use, the present invention enables intelligent voice robots equipped with the present invention to possess certain learning and judgment abilities, thereby better serving users.

[0084] Furthermore, during use, this invention can periodically and autonomously update the recommended items, thereby ensuring that the intelligent voice robot equipped with this invention keeps pace with the times and better suits users' habits. At the same time, the top-level control logic program in this invention can assist the invention in providing multiple recommendation schemes for different groups of people, thus further improving the compatibility of this invention.

[0085] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An improved user similarity based collaborative filtering recommendation algorithm, characterized in that, Comprising the following steps: Step1: Obtain the task group target, judge the task group target type, and analyze the task group target type corresponding to the use scenario; Step2: According to the task group target fault setting result, the task group target is classified, and the reference task target group classification is used to control the tone, timbre and loudness of the voice assistant program in the hardware; Step3: The control results of the tone, timbre and loudness of the voice assistant program in the hardware are matched to the corresponding use scenario; Step4: Periodically collect user demand data, and evaluate the user attributes, corresponding use scenarios and voice assistant related control matching degree according to the collected data; Step5: Set the matching degree evaluation qualified index, and capture the user similar operation behavior characteristics after confirming that the matching degree evaluation meets the evaluation qualified index; Step6: According to the captured user similar operation behavior characteristics, generate user demand prediction items, send the user demand prediction items to the initialization recommendation items and top them; Step7: The user sets the detection period to count and calculate the frequency of use of the user demand prediction items on the top, and sets the use index; Step8: According to the set use index, the user demand prediction items are screened, and the screened items are recorded as user demand optimization recommendation items; Step9: Obtain the highest recommendation item of the initialization recommendation item under the top user demand prediction item, set the highest recommendation item as the optimization recommendation item, and replace the user demand prediction item with the optimization recommendation item to top, and set the top control logic program. 2.The collaborative filtering recommendation algorithm based on improved user similarity according to claim 1, wherein, The number of use scenario items corresponding to the task group target type analyzed in step 1 is greater than or equal to 1, wherein the number of use scenario items is a natural number. 3.The improved user similarity-based collaborative filtering recommendation algorithm of claim 1, wherein, The step Step1 is provided with a sub-step, Comprising the following steps: Step11: Collect the task group target demand content and collect the task group target attributes; Step12: According to the collected task group target attribute, set the task group target fault; Wherein, step Step11 and step Step12 are specifically used for step Step1 in task group target type judgment and corresponding use scenario.

4. The collaborative filtering recommendation algorithm based on improved user similarity according to claim 1, characterized in that, The control of the tone, timbre and loudness of the voice assistant program in the hardware in step Step2 also includes the use scenario corresponding to the task group target type analyzed in step Step1.

5. The collaborative filtering recommendation algorithm based on improved user similarity according to claim 1, characterized in that, The user demand data collection period in step Step4 is set to be less than or equal to seven days.

6. The collaborative filtering recommendation algorithm based on improved user similarity according to claim 1, characterized in that, The user demand initialization recommendation item in step Step4 is initially set according to the use scenario corresponding to the task group target type analyzed in step Step1 and the task group target fault area division result set in step Step12.

7. The collaborative filtering recommendation algorithm based on improved user similarity according to claim 1, characterized in that, Step5 is executed again in the state that the matching degree evaluation does not meet the evaluation qualified standard, and the processing result of step Step11 is sent to step Step12, and step Step12 re-calls the target user in each classification of the task group target according to the transmission content of step Step11. 8.The improved user similarity based collaborative filtering recommendation algorithm of claim 1, wherein, The statistical logic in the step Step7 is the ratio between the use frequency of each item in the user demand prediction item and the user quantity; the calculation logic is the ratio between the use frequency of all items in the user demand prediction item and the user demand prediction item delivery days.

Images