Cabin function index evaluation mechanism construction method fusing user emotional tendency correction

Abstract

The invention discloses a cabin function index evaluation mechanism construction method fusing user emotional tendency correction, and the main design concept is that the emotional tendency in user evaluation is corrected; and establishing a multi-level evaluation strategy with a unified benchmark for user emotion tendency correction, product function dimension empowerment and user experience evaluation index empowerment, and realizing fine management of user evaluation data. Specifically, the method abandons the idea of optimization from an equipment side, and finds another way from a user side and fully considers individual differences of users: 1, screening high-quality evaluation individuals and evaluation data thereof from a large sample, and endowing the high-quality evaluation individuals and the evaluation data with a relatively high weight, so that the professional and credible evaluation result is ensured; secondly, non-objective user evaluation data with obvious emotional tendency are compensated through an emotional correction means, and evaluation scores are pulled back to a datum line in a weight compensation setting mode; the objectivity of the evaluation data of the intelligent cabin of the vehicle is effectively improved in a low-cost mode.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, and in particular to a method for constructing a cockpit function indicator evaluation mechanism that integrates user sentiment correction. Background Technology

[0002] Compared to traditional cars, intelligent connected vehicles have added numerous interactive modules within the cabin, giving rise to the concept of intelligent cockpits. With the further integration of IoT technology and AI big data models into vehicles, the focus of the automotive experience will shift further towards the cockpit. Cockpit functions such as intelligent voice interaction, vehicle-to-everything (V2X) connectivity, and microphone-free karaoke are becoming increasingly abundant. However, whether this accumulation of functions truly delivers a better user experience, and how it affects user evaluation of the cockpit experience, requires further research. Existing cockpit evaluation methods typically focus on the feasibility of functions, the smoothness of interaction, and ease of use, often resulting in higher evaluations for more functions. Furthermore, evaluation results are easily influenced by users' life experiences and emotional preferences, often resulting in subjective biases and deviations from the actual competitiveness of the vehicle's cockpit. Therefore, existing cockpit evaluation methods cannot meet the needs of companies in assessing their product's competitiveness against competitors, hindering product experience optimization.

[0003] Existing methods for evaluating cabin user experience mainly include two types: small-sample engineer evaluations (usually around 5 people) and large-sample Beta user evaluations (usually involving online public opinion surveys with dozens or hundreds of participants). Engineer evaluations differ from actual user experiences due to their different environment and user backgrounds. Therefore, user evaluations are essential. However, current user evaluations do not differentiate between users, and their results are often simply averages, leading to a lack of data rigor.

[0004] Currently, large-sample evaluation methods mainly follow the traditional data statistical analysis based on online user reviews used in internet products. However, the automotive industry differs from other industries; intense competition among companies vying for market share leads to significant uncertainty regarding the objectivity of online reviews. Even though more and more companies are incorporating beta testing into their user experience evaluation processes, while the evaluation results are abundant, they cannot avoid negative impacts such as inconsistent evaluation standards, varying evaluation quality, and the dominance of user subjectivity. Therefore, optimizing big data for smart cockpit user experience evaluations, such as minimizing subjective biases and increasing objectivity and credibility, is one of the main problems to be solved in the current cockpit development stage. Summary of the Invention

[0005] In view of the above, the present invention aims to provide a method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction, so as to solve the aforementioned technical problems.

[0006] The technical solution adopted in this invention is as follows:

[0007] This invention provides a method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction, including:

[0008] User types are defined in advance based on user feedback data;

[0009] Based on large sample evaluation data, flexible correction amounts are set for different types of users to correct evaluation sentiment.

[0010] The functions of the smart cockpit are broken down and set into multiple dimensions, with different weights assigned to different dimensions of different functions.

[0011] Several target evaluation indicators for measuring the smart cockpit experience were selected and weighted accordingly.

[0012] By combining the correction of evaluation sentiment, the multi-level weighting of intelligent cockpit functions, and the target evaluation index system, a cockpit function index evaluation mechanism is formed.

[0013] In at least one possible implementation, the elasticity correction amount is calculated in the following ways:

[0014] Within a preset time span, collect multi-round evaluation data from different types of users on several tested vehicles;

[0015] The initial evaluation data is compared with the final evaluation data, and the magnitude of the evaluation change is calculated as the elastic adjustment amount for different types of users.

[0016] In at least one of the possible implementations, the elasticity correction amount is calculated using the following formula:

[0017] The elastic adjustment amount is calculated as (S2-S1) / S1, where S1 is the initial evaluation score of the user of the corresponding type, and S2 is the final evaluation score of the user based on the time span.

[0018] In at least one possible implementation, the user type classification includes:

[0019] Use preset scoring rules to score user feedback data and determine user tags;

[0020] User types are clustered based on user tags.

[0021] In at least one of the possible implementations, setting different weights for different dimensions of different functions includes: using a unified AHP (Analytic Hierarchy Process) method for different cockpit function levels and determining the weights of each level in combination with expert experience.

[0022] In at least one of the possible implementations, the evaluation mechanism construction method further includes: after setting the elasticity adjustment amount, constructing a random forest automatic classification model, automatically matching the corresponding elasticity adjustment amount according to the defined user type, and automatically correcting the sentiment tendency of the user evaluation data using the corresponding elasticity adjustment amount.

[0023] Compared with existing technologies, the main design concept of this invention lies in proposing a correction of sentiment bias in user reviews and constructing a multi-level evaluation strategy with a unified benchmark, which includes user sentiment bias correction, product function dimension weighting, and user experience evaluation index weighting, thereby achieving refined management of user review data. Specifically, this invention abandons the approach of optimizing from the device side and instead takes a different approach from the user side, fully considering individual differences among users: First, it selects high-quality individual reviewers and their review data from a large sample and assigns them higher weights, thereby ensuring the professional credibility of the evaluation results; Second, it compensates for non-objective user review data with obvious sentiment bias through sentiment correction, using a weighted compensation method to bring the evaluation score back to the baseline. This invention effectively improves the objectivity of vehicle intelligent cockpit evaluation data in a low-cost manner. Attached Figure Description

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described below with reference to the accompanying drawings, wherein:

[0025] Figure 1 A schematic diagram illustrating the construction method of the cockpit function index evaluation mechanism that integrates user emotional tendency correction provided in the embodiments of the present invention;

[0026] Figure 2 This is a schematic diagram of the vehicle intelligent cockpit evaluation index system provided in an embodiment of the present invention. Detailed Implementation

[0027] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0028] This invention proposes an embodiment of a method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction. Specifically, as follows: Figure 1 As shown, it includes:

[0029] Step 10: Determine user types in advance based on user feedback data;

[0030] For example, before evaluating the cabin experience, users can fill out a pre-made survey questionnaire with their consent and in compliance with laws and regulations. This can collect information that is strongly related to intelligent connected vehicles, quantify users' familiarity with new energy vehicles through the questionnaire results, determine the value type of user evaluations, and further determine the decisive role of user evaluations in the overall vehicle results. This can help select high-quality users, filter out low-quality users, and reduce the weakening effect of low-quality evaluations on the objectivity of the overall evaluation results.

[0031] The questionnaire content is not the focus of this invention; the table below is for illustrative purposes only:

[0032]

[0033] The feedback data from the questionnaires were analyzed using the Analytic Hierarchy Process (AHP) combined with expert scoring (the method is the same as the AHP process described below). The final calculation method is as follows:

[0034] User value total score S = (Score of Question 1 + Score of Question 2 + Score of Question 3) * Familiarity dimension weighting + (Score of Question 4 + Score of Question 5) * Attitude dimension weighting.

[0035] Users are scored and tagged using preset scoring rules combined with user feedback data, thus segmenting them for subsequent weighting based on their impact on cabin evaluation results. In some embodiments, user types are clustered into four categories based on tag clustering, using two core dimensions: experience level and attitude tendency.

[0036] The first type of user is the expert-level high-value user: familiar with new energy technologies, long-term user of intelligent cockpit functions, willing to try new technologies, and provides professional and detailed feedback.

[0037] The second type of user is characterized by eager early adopters: they have a strong interest in smart technology but limited experience (e.g., they have only briefly experienced it), and may give subjective evaluations due to their novelty.

[0038] The third type of user is characterized as neutral and pragmatic: they have some experience, but hold a pragmatic attitude towards intelligent functions (such as "good enough is fine"), and their evaluations are relatively rational.

[0039] The fourth type of user is characterized by a negative and conservative attitude: they distrust intelligent functions (such as concerns about privacy and security) or lack of experience in using them, which leads to a large bias in their evaluation.

[0040] For information on the two core dimensions mentioned above, please refer to the table below:

[0041]

[0042] The first type of user: The total score S≥S1 and the familiarity score E≥E1, which is suitable for users with high total score and experience score;

[0043] The second type of user: S1 > total score S≥S2, familiarity level E1 > E≥E2 and T1 > attitude score ≥T2, suitable for users with average experience and average attitude.

[0044] The third type of user: S2 > total score S, familiarity level E < E2 and attitude score ≥ T1, suitable for users with high attitude score and low experience score;

[0045] Passive and conservative users: S2 > total score S, familiarity level E < E2 and attitude score < T2, suitable for users who lack experience and have low attitude scores.

[0046] Step 20: Based on large sample evaluation data, set flexible correction amounts for different types of users to correct evaluation sentiment.

[0047] To ensure that the evaluation focuses on the product itself, the sentiment of the user evaluation data from the aforementioned categories needs to be adjusted accordingly. This is to avoid subjective and negative user evaluation data interfering with the evaluation results. Please refer to the table below:

[0048]

[0049] The calculation method for the elastic correction amount may include:

[0050] Within a preset time span, collect multi-round evaluation data from different types of users on several tested vehicles;

[0051] The initial evaluation data is compared with the final evaluation data, and the magnitude of the evaluation change is calculated as the elastic adjustment amount for different types of users.

[0052] To illustrate the method for determining the value of the elastic correction amount, the above steps will be explained with examples:

[0053] a. First, recruit 10 users who are eager to try new things, 10 users who are neutral and practical, and 10 users who are passive and conservative. Prepare 8 car models with the same level of cabin product experience and randomly divide them into four groups.

[0054] b. Conduct a four-month (M1-M4) evaluation activity, inviting the above users to conduct an evaluation once a month, and each time conduct an experience test on randomly grouped vehicles.

[0055] c. Evaluate a different group of vehicles one month later, and repeat this cycle three times.

[0056] d. Collect the scoring data from the four evaluations, compare the first scoring result S1 with the fourth result S2, calculate the average change range, and determine the elasticity correction amount.

[0057] The aforementioned calculation method, which uses adjustments based on users' prior and subsequent experiences, takes the fourth type of user as an example. After gaining some experience, their evaluation results may revert to an objective level. Therefore, their elastic adjustment amount can be calculated using the following formula:

[0058]

[0059] Among them, S1 is the rating given by the fourth type of user when experiencing a certain model for the first time, and S2 is the rating given by the user after four months of familiarizing themselves with the new energy cabin.

[0060] Furthermore, after obtaining the elasticity adjustment amount for different types of users through the above embodiments, a random forest automatic classification model can be constructed to realize the complete process from user survey questionnaire input (judging the user value type), to automatic matching of elasticity adjustment amount, and then to automatic correction calculation of evaluation results.

[0061] Step 30: Deconstruct the functions of the smart cockpit and set multiple dimensions, and assign different weights to different dimensions of different functions;

[0062] In other words, to evaluate the multi-level product user experience level of each function of the smart cockpit, it is first necessary to clarify the evaluation function and assign functions based on the principle of their decisive influence on the overall intelligence level of the cockpit.

[0063] In some preferred embodiments of the present invention, three levels of dimensions are set for each cockpit function. For example, the first level includes four types: "perception capability," "interaction capability," "service capability," and "connectivity capability," and the sum of the weights of the four is 1. Users value the perception capability in the cockpit more, as it has a greater impact on users' purchasing decisions. Therefore, the perception capability module is given a higher weight. For example, in one embodiment, the first-level function "perception capability" is weighted at 35%. Similarly, "interaction capability" is weighted at 35%, "service capability" at 15%, and "connectivity capability" at 15%. This is the weighting of the first level dimension.

[0064] The second-level dimensions are the breakdown of each first-level dimension. For example, the second-level dimensions of "perception ability" include three types: driver status monitoring (weighted 30%), driver identification (weighted 30%), and environmental perception and driving information display (weighted 40%). Similarly, the weight of the three is 1.

[0065] The third-level dimension is a breakdown of each of the second-level dimensions. For example, the third-level dimension of "driver identification" includes two types: physiological feature recognition (weighted at 40%) and digital key (weighted at 60%). Similarly, the weight of the two is 1.

[0066] Therefore, by decomposing and assigning weights to functions at each level, the goal of achieving highly granular and comprehensive evaluation can be achieved.

[0067] The cockpit function weighting method proposed in this invention preferably adopts the AHP (Analytic Hierarchy Process) method uniformly, and is determined in conjunction with expert experience, as follows:

[0068] A. Establish a hierarchical structure model. First, construct the intelligent cockpit experience evaluation system into a hierarchical structure model, including an objective layer, a criterion layer, and a solution layer. Objective layer: Overall user experience level of the intelligent cockpit. Criterion layer: Includes perception capabilities, interaction capabilities, service capabilities, and interconnectivity capabilities. Solution layer: The secondary dimensions under the criterion layer, such as driver status monitoring and passenger status monitoring under driver status monitoring; and central control screen interaction, auxiliary screen interaction, and steering wheel interaction under touch interaction.

[0069] B. Constructing the Judgment Matrix. Five experts in the field of intelligent cockpits (including cockpit design, subjective evaluation, etc.) were invited to construct a judgment matrix by comparing each factor in pairs at the same level using the Saaty 1-9 scale, as shown in the following figure.

[0070]

[0071] C. Hierarchical Single Ranking and Consistency Test. Calculate the maximum eigenvalue λmax and the corresponding eigenvector for the judgment matrix. Normalizing the eigenvectors yields the weight distribution of each factor in the previous level. Simultaneously, to avoid inconsistencies in expert judgments, a consistency test is required. The test steps are as follows: ① Calculate the consistency index; ② Look up the corresponding RI value based on the matrix order; ③ Calculate the consistency ratio, i.e., the CR value. When the CR value < 0.1, the consistency test is satisfied.

[0072] D. Overall Hierarchical Ranking and Consistency Test. The steps are the same as for single-level hierarchical ranking consistency, calculating the ranking weights of all factors at the same level relative to the highest level.

[0073] Step 40: Select several target evaluation indicators to measure the smart cockpit experience and assign weights to each evaluation indicator;

[0074] In practice, metrics for measuring cabin user experience can be selected from perspectives such as user experience elements, user experience quantification methods, and product usability. This evaluation system unfolds from three primary dimensions: product functionality, product performance, and user experience evaluation. Following the step-by-step breakdown of cabin functions described earlier, secondary and tertiary dimensions can be set under each primary dimension. Figure 2 This is an illustration (for reference only, not as a limitation, and has no necessary impact on the overall concept). The construction approach may include:

[0075] a. The product has functions, and many functions. In terms of the richness of functions, generally speaking, the more functions a product has, the more satisfied users will be.

[0076] b. In addition to having functionality, the functionality performs well. This mainly involves the stability of the functionality, the speed of operation, and the success rate of function execution, focusing on the objective performance of the product.

[0077] c. Based on good functional performance, assess user satisfaction. This mainly involves operational convenience and psychological satisfaction, focusing on the user's emotional experience.

[0078] Step 50: Combining the correction of the evaluation sentiment, the multi-level weighting of intelligent cockpit functions, and the target evaluation index system, a cockpit function index evaluation mechanism is formed.

[0079] In summary, the main design concept of this invention lies in proposing a correction of sentiment bias in user evaluations and constructing a multi-level evaluation strategy with a unified benchmark, encompassing user sentiment bias correction, product function dimension weighting, and user experience evaluation index weighting, thereby achieving refined management of user evaluation data. Specifically, this invention abandons the approach of optimizing from the device side, instead taking a different approach from the user side and fully considering individual user differences: First, it selects high-quality individual evaluators and their evaluation data from a large sample and assigns them higher weights, thus ensuring the professional credibility of the evaluation results; second, it compensates for non-objective user evaluation data with obvious sentiment bias through sentiment correction methods, using a weighted compensation approach to bring the evaluation scores back to the baseline. This invention effectively improves the objectivity of vehicle intelligent cockpit evaluation data in a low-cost manner.

[0080] In this invention, when directional terms are mentioned, they are relative concepts based on the embodiments. Furthermore, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects have an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.

[0081] The above description of the structure, features, and effects of the present invention is based on the embodiments shown in the figures. However, the above are only preferred embodiments of the present invention. It should be noted that the technical features involved in the above embodiments and their preferred methods can be reasonably combined and matched by those skilled in the art to form a variety of equivalent solutions without departing from or changing the design concept and technical effects of the present invention. Therefore, the present invention is not limited to the scope of implementation shown in the figures. Any changes made in accordance with the concept of the present invention, or modifications to equivalent embodiments, that do not exceed the spirit covered by the specification and figures, should be within the protection scope of the present invention.

Claims

1. A method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction, characterized in that, include: User types are defined in advance based on user feedback data; Based on large sample evaluation data, flexible correction amounts are set for different types of users to correct evaluation sentiment. The functions of the smart cockpit are broken down and set into multiple dimensions, with different weights assigned to different dimensions of different functions. Several target evaluation indicators for measuring the smart cockpit experience were selected and weighted accordingly. By combining the correction of evaluation sentiment, the multi-level weighting of intelligent cockpit functions, and the target evaluation index system, a cockpit function index evaluation mechanism is formed.

2. The method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction according to claim 1, characterized in that, The calculation method for the elastic correction amount includes: Within a preset time span, collect multi-round evaluation data from different types of users on several tested vehicles; The initial evaluation data is compared with the final evaluation data, and the magnitude of the evaluation change is calculated as the elastic adjustment amount for different types of users.

3. The method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction according to claim 2, characterized in that, The elastic correction amount is calculated using the following formula: The elastic adjustment amount is calculated as (S2-S1) / S1, where S1 is the initial evaluation score of the user of the corresponding type, and S2 is the final evaluation score of the user based on the time span.

4. The method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction according to claim 1, characterized in that, The defined user types include: Use preset scoring rules to score user feedback data and determine user tags; User types are clustered based on user tags.

5. The method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction according to claim 1, characterized in that, The setting of different weights for different dimensions of different functions includes: using a unified AHP (Analytic Hierarchy Process) method for different cockpit function levels, and combining expert experience to determine the weights of each level.

6. The method for constructing a cockpit function index evaluation mechanism that integrates user emotional tendency correction according to any one of claims 1 to 5, characterized in that, The evaluation mechanism construction method further includes: after setting the elasticity adjustment amount, constructing a random forest automatic classification model, automatically matching the corresponding elasticity adjustment amount according to the defined user type, and automatically correcting the sentiment tendency of the user evaluation data using the corresponding elasticity adjustment amount.

Images