An interface layout recommendation method for intelligent devices

Abstract

The application discloses a kind of interface layout recommendation method for intelligent device, it is related to interface layout technical field, including the following steps: obtaining user behavior data, and according to user behavior data, user behavior model is established;Reflecting user current use habit and the user time sequence behavior model of preference is constructed;According to the importance score of each interface element using weighted integral algorithm calculation multidimensional data in user behavior data;According to the importance score of interface element, page control is reordered and restructured;Combining the font size use behavior of user's history, the font size of each interface element is adjusted, the font size of interface element of importance score high increases, and the font size of interface element of importance score low shrinks;The interface of the application can be dynamically adjusted to adapt to the individual needs of each user.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of interface layout, in particular to an interface layout recommendation method for a smart device. BACKGROUND

[0002] The interface design of existing smart devices is often designed according to average user needs, lacking truly personalized recommendations for each user's physiological characteristics and usage habits.

[0003] That is, the traditional smart device interface usually only provides preset parameter adjustment (such as font size, color contrast, etc.) or fixed layout, and fails to dynamically optimize the interface according to real-time feedback or behavior of the user; secondly, many interface design methods still use static layout, which cannot perceive and dynamically adjust according to the user's operation behavior (such as click frequency, touch force, reaction speed, etc.), which limits the improvement of interaction efficiency. SUMMARY

[0004] To solve the defects in the prior art, the present application provides an interface layout recommendation method for a smart device.

[0005] To solve the above technical problems, the present application provides the following technical solutions:

[0006] The present application provides an interface layout recommendation method for a smart device, comprising the following steps:

[0007] Obtain user behavior data, the user behavior data including historical behavior data and real-time operation data, the user behavior data including user click operation coordinates, executed interaction types, each interaction dwell time, current task operation intention types, and interaction operation result identifiers, and establish a user behavior model according to the user behavior data;

[0008] Map the user behavior data to a structured behavior vector, and fuse operation behaviors at different time points, adjust the behavior weight of each behavior record according to a time decay factor, and construct a user time sequence behavior model reflecting the user's current usage habits and preferences;

[0009] According to the multi-dimensional data in the user behavior data, a weighted integral algorithm is used to calculate the importance score of each interface element, and the weight of each dimension is adjusted based on a user portrait update strategy;

[0010] According to the importance score of the interface element, the page controls are reordered and reconstructed, the interface elements with high scores are preferentially presented in the main visual area, and the controls with the same score are further sorted according to the historical hot area priority;

[0011] In combination with the historical font size usage behavior of the user, the font size of each interface element is adjusted, the font size of the interface element with high importance score is increased, and the font size of the interface element with low importance score is reduced.

[0012] In each interface reordering and reconstruction process, an animation transition mechanism based on the current change amplitude and user historical cognitive load is adopted to realize smooth transition of the interface change process.

[0013] As a preferred technical solution of the present application, the user behavior model is constructed as follows:

[0014] ;

[0015] Wherein, n represents the number of user operation records;

[0016] w i represents the time decay weight of the i-th behavior record, and is represented as:

[0017] ;

[0018] Wherein, λ is a parameter factor controlling the decay degree of historical behavior influence, T now is the current system timestamp, T i is the timestamp of the i-th behavior record;

[0019] u i represents the feature vector of the i-th operation behavior, and is represented as:

[0020] ;

[0021] Wherein, p i is the click position represented by two-dimensional coordinates, t i is the operation type, d i is the element dwell time, a i is the operation intention category, and r i is the operation result.

[0022] As a preferred technical solution of the present application, the importance score of the interface element is determined by the following formula:

[0023] ;

[0024] Wherein, E j represents the interface element, S(E j ) represents the importance score of the interface element, f click (E j ) is the click frequency of the element, f stay (E j) is the average time of the user staying in the element, f path (E j ) is the frequency of the element in the user access path, f success (E j ) is the success rate of the element operation, f intent (E j ) is the operation intensity index related to the current interaction intention, and alpha, beta, gamma, delta and epsilon are weight coefficients.

[0025] As a preferred technical solution of the application, the reordering and reconstruction of the page control adopts a multi-objective optimization sorting algorithm based on stability constraints, taking the importance score of the interface element as the main sequence index, combining historical layout inertia and user familiarity to construct a target function:

[0026] ;

[0027] Wherein, V(E j ) represents the target function score, F(E j ) represents the user familiarity, which is obtained by weighting the interaction times and the stay time of the user to the control, D(E j ) represents the displacement distance of the current position information of the control and the last layout, and eta1, eta2 and eta3 are weight coefficients.

[0028] As a preferred technical solution of the application, the reordering and reconstruction of the page control follows a joint optimization strategy of reachability and visual focus, and the steps are as follows:

[0029] According to the importance score of the interface element, all controls are divided into priority intervals, including high score zone, intermediate score zone and low score zone;

[0030] The controls in the high score zone are distributed in the hot area;

[0031] The controls in the intermediate score zone are distributed in the secondary hot area;

[0032] The controls in the low score zone are concentrated in the edge area.

[0033] As a preferred technical solution of the application, the adjustment of the font size is based on the deviation of S(E j ) from the mean value mu of the scores of all interface elements, and a magnification mechanism is implemented by standard deviation sigma, and the specific formula is as follows:

[0034] ;

[0035] Wherein, mu represents the mean value of the scores of all interface elements, sigma is the standard deviation, and theta is the magnification coefficient.

[0036] As a preferred technical solution of the application, the animation transition mechanism is driven by a cognitive load model, and the animation duration T anim is determined according to the following formula:

[0037]

[0038] wherein T0 is a basic minimum animation duration, C hist represents the average cognitive load of the user history, C δ represents the total visual displacement of the control caused by this rearrangement, and p is the response coefficient.

[0039] As a preferred technical solution of the application, the interface layout recommendation method further comprises the following steps:

[0040] After the interface adjustment is completed, automatically generate voice prompt information, and point out the changed area or control in the interface through voice broadcast, and the prompt content includes the control name and change type.

[0041] As a preferred technical solution of the application, the voice prompt information is generated by a Top-K change area selection strategy, specifically:

[0042] Calculate the change degree of all controls in the interface;

[0043] Sort and select the top n controls in the change degree;

[0044] Generate a voice template prompt for each control.

[0045] The application has the following advantages:

[0046] 1. In the application, the behavior data (such as click frequency, dwell time, operation intention, etc.) of the user is collected and analyzed in real time, and the interface can be dynamically adjusted to adapt to the individual needs of each user. Unlike traditional interfaces that rely only on preset adjustments (such as font size, contrast, etc.), this method builds a user behavior model and combines time sequence weights to optimize the interface layout in real time, ensuring that the interface design is more in line with the user's operation habits and needs. This dynamic adaptability significantly improves the user's operation efficiency and reduces the occurrence of operation errors.

[0047] 2. In the application, through interface rearrangement, font size adjustment, animation transition and other methods, the user's cognitive burden is reduced while ensuring smooth interface changes, for example, the interface rearrangement based on user historical behavior preferentially displays high importance controls, reducing the difficulty of users finding target functions.

[0048] ​3.The interface layout recommendation method of the present application can optimize the layout order of the controls on the basis of ensuring interface familiarity, ensure that important function areas are presented in the main visual area of the screen in priority, and further improve the accessibility and interaction efficiency of the interface through intelligent adjustment of the layout and position of the controls according to the operation frequency and habits of the user, which is particularly suitable for user groups to make them more convenient and efficient when using the interface. BRIEF DESCRIPTION OF DRAWINGS

[0049] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate the present application, and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation on the present application. In the drawings:

[0050] Figure 1 The flowchart of the interface layout recommendation method of the present application. DETAILED DESCRIPTION

[0051] The preferred embodiments of the present application are described below in conjunction with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and do not constitute a limitation on the present application.

[0052] As shown in the drawings, an interface layout recommendation method for a smart device includes the following steps: Figure 1

[0053] Obtain user behavior data, the user behavior data including historical behavior data and real-time operation data, the user behavior data including the click operation coordinates of the user, the interaction type performed, the dwell time of each interaction, the operation intention type of the current task, and the result identifier of the interaction operation, and establish a user behavior model according to the user behavior data;

[0054] Map the user behavior data to a structured behavior vector, fuse the operation behaviors at different time points, dynamically adjust the behavior weight of each behavior record according to a time decay factor, and construct a user time sequence behavior model reflecting the current use habits and preferences of the user;

[0055] According to the multi-dimensional data in the user behavior data, the importance score of each interface element is dynamically calculated by using a weighted integral algorithm, and the weight of each dimension is adaptively adjusted based on a user portrait update strategy;

[0056] According to the importance score of the interface element, the page controls are reordered and reconstructed, the interface elements with high scores are presented in the main visual area in priority, and the controls with the same scores are further sorted according to the historical hot area priority to improve the accessibility and visibility of the main function module;

[0057] ​The font size of each interface element is dynamically adjusted according to the visual preference settings and historical font size usage behaviors of the user, the font size of the interface element with a high importance score is increased, and the font size of the interface element with a low importance score is reduced, so that the visual burden of the user is reduced.

[0058] In each interface reordering and reconstruction process, an animation transition mechanism based on the current change amplitude and the historical cognitive load of the user is adopted to realize smooth transition of the interface change process.

[0059] The core technical scheme of the application is to collect the behavior characteristics and interaction data of the user, and dynamically adjust the interface display mode, so as to realize interface optimization more in line with the user's needs. Specifically, the method combines various user behavior data such as click coordinates, interaction types, stay time, operation intention types and operation results, generates a behavior model reflecting the user's operation habits by analyzing these data, and through the model, the application can adjust the interface layout in real time according to the user's behavior characteristics, so as to improve the user's operation efficiency and interface understandability.

[0060] It should be noted that the interface elements and controls in the application refer to the same concept, both refer to visual components or modules that the user can interact with in the interface.

[0061] In the intelligent device interface layout recommendation method of the application, a stable preference control mechanism is further included, which mainly sets a stable preference control field StableMode in the user portrait, used to indicate whether the user enables the interface dynamic adjustment function. The field is initialized and configured by the system, and the user can manually adjust it in the system settings.

[0062] The specific working mechanism includes:

[0063] StableMode field setting method:

[0064] StableMode is a Boolean value, True indicates that the user opens the stable preference mode and closes the interface dynamic reconstruction, False indicates that dynamic interface adjustment is allowed according to the user behavior model, and the system is initially configured as True by default.

[0065] Interface reconstruction trigger condition control:

[0066] When StableMode=True, the system completes the importance scoring of the interface elements according to the user behavior model, but does not directly trigger the reordering or font size adjustment of the controls, only records and caches the recommended results. When the user explicitly selects "open intelligent optimization recommendation" in the settings interface, StableMode is set to False, and the system dynamically updates the interface layout according to the behavior model.

[0067] Guidance and prompt mechanism:

[0068] When the user is first prompted to enable the dynamic recommendation function, the system explicitly informs the user of the possible interface changes through the interface prompt method and prompts the user that the function can be turned off at any time. Before each layout change operation, the system uses a change amplitude judgment mechanism based on the user's historical cognitive vector to trigger an animation transition interface adjustment only when the change amplitude exceeds the preset threshold, and provides a voice or visual prompt for the adjustment content.

[0069] In this process, the above-mentioned stable preference control mechanism works with the interface reordering mechanism and the animation transition mechanism to ensure that the interface recommendation is both intelligent and takes into account the user's demand for information stability, achieving the goal of "user controllable + system intelligent".

[0070] Further, the user behavior model is constructed as follows:

[0071] ;

[0072] where n represents the number of user operation records;

[0073] w i represents the time decay weight for the i-th behavior record, which is represented as:

[0074] ;

[0075] where λ is a parameter factor that controls the degree of influence of historical behavior decay, which is a positive number, T now is the current system timestamp, T i is the timestamp of the i-th behavior record;

[0076] u i represents the feature vector of the i-th operation behavior, which is represented as:

[0077] ;

[0078] where p i is the click position represented by two-dimensional coordinates, t i is the operation type (such as click, swipe, long press, etc.), d i is the element dwell time, reflecting the user's attention allocation, a i is the operation intent category (such as search, confirmation, input, etc.), which is inferred through task context and sequence modeling, r i is the result of this operation, with success recorded as 1 and failure recorded as 0.

[0079] This formula is designed to solve two core technical problems in user behavior modeling:

[0080] Dynamicity: User's interaction behavior changes over time, early behavior's value for current interface optimization decreases gradually;

[0081] Multi-dimension behavior fusion: Single dimension (e.g. click times) is not enough to reflect user's real usage habits, position, duration, intent and operation success rate must be considered comprehensively.

[0082] By introducing w i , the model can balance the influence of recent behavior and historical behavior. For example, a user frequently clicks a certain button last week, but almost never uses it this week, then the importance of the button will naturally decrease over time.

[0083] The user behavior model builds a behavior model that comprehensively reflects user operation habits and preferences through analysis of historical behavior data and real-time operation data. The behavior weight is dynamically adjusted using λ because user behavior changes over time, especially for users, long time without performing a certain operation may cause the behavior to gradually decrease its influence on them. Therefore, using time decay can better reflect the user's current behavior habits.

[0084] At work, first, the front-end collection SDK acquires the following interaction data in real time through DOM binding events: two-dimensional coordinate representation of click position, operation type, element dwell time, operation intent, operation result. The data is structured as a behavior log, and each log corresponds to the above feature vector u i .

[0085] For each log, the decay weight is calculated according to the time difference between the current time and the behavior record. Preferably, λ takes the value range: [0.001, 0.05]. If λ is too large, the system relies too much on recent behavior; if it is too small, it relies too much on historical behavior. For mobile interaction interfaces, λ = 0.02 is more balanced, which can effectively reflect the user behavior decay trend of the past week.

[0086] Finally, the model H represents a multi-dimensional weighted average behavior state:

[0087] If a user frequently clicks a certain control recently, the distribution of p i will be concentrated in that area;

[0088] If the user often uses "confirmation" type interactions, the probability distribution of a i will be biased towards "confirmation";

[0089] Elements with high dwell time d i represent high user attention.

[0090] Model H is used as the input basis for subsequent layout rearrangement and font size adjustment.

[0091] Further, the importance score of the interface element is determined by the following formula:

[0092] ;

[0093] Wherein, E j represents the interface element, S(E j ) represents the importance score of the interface element, f click (E j ) is the click frequency of the element, which represents the activity of being operated by the user, and the calculation method is: counting the number of times that the user clicks E j in a certain time window (such as 7 days or 30 days) and normalizing it to a frequency value;

[0094] f stay (E j ) is the average time that the user stays on the element, reflecting the user's attention, and the calculation method is: taking the average value of all interaction stay times on E j and normalizing it;

[0095] f path (E j ) is the frequency of the element in the user access path, which represents the key degree in the task flow, and the calculation method is: counting the frequency of E j appearing in the task behavior sequence (i.e. the number of times appearing in the task path);

[0096] f success (E j ) is the success rate of the element operation, which is used to evaluate the ease of use of its function, and the calculation method is: calculating the ratio of the number of successes to the total number of attempts;

[0097] f intent (E j ) is the operation intensity index related to the current interaction intention, which is used to measure its context relevance, by extracting the current task intention I t through the model, combining historical data to analyze the relevance of the control E j and I t , which is represented as follows:

[0098] ;

[0099] Wherein, represents the conditional probability of the control E j being used under the intention I t ;

[0100] α, β, γ, δ, ε are weight coefficients respectively, and satisfy α+β+γ+δ+ε=1.

[0101] α (click frequency weight): The recommended setting range is [0.2, 0.35], and the click is a significant sign of behavior, which should be given a medium or higher weight;

[0102] β (residence time weight): Recommended as [0.1, 0.25], because it is affected by the difference in user reading speed, the weight is slightly lower;

[0103] γ (path frequency weight): Suggested to be set in [0.15, 0.3], reflecting the role of the control in completing the task;

[0104] δ (success rate weight): Suggested to be set as [0.1, 0.2], used to shield low-quality controls;

[0105] ε (intention matching weight): Should be dynamically adjusted, when the intention recognition accuracy is high, it can be set to a high value [0.2, 0.35] to enhance the adaptability of the system.

[0106] The recommended initial default values are as follows (for general users):

[0107] α=0.25 (click frequency dominates), β=0.20, γ=0.20, δ=0.15, ε=0.20.

[0108] For example: the user repeatedly clicks "affairs guide" but the success rate is low, the system adjusts f intent and f success , significantly improves the position of the control, and increases the prompt animation to reduce misoperation.

[0109] By weighting the scores of multiple dimensions (click frequency, residence time, access path frequency, operation success rate, and operation intention strength), the comprehensive importance score of each interface element is obtained. The weighted integral algorithm is used because different dimensions of behavior characteristics have different effects on user operation habits and interface use experience. For example, elements with high click frequency indicate that users have strong demand for the elements, while elements with low operation success rate may need to be adjusted to improve user experience. In this way, the importance of each interface element can be accurately evaluated, and reasonable layout optimization can be performed.

[0110] It should be noted that in actual use, controls may sometimes be misjudged as "negligible" elements due to low interaction frequency. In order to further enhance the executability and operation intuition of user key tasks, especially for user groups with declining cognitive or visual abilities, a semantic importance indicator is added to the original interface element importance score to introduce a semantic priority factor, i.e., to judge the importance of semantics.

[0111] The semantic importance index modeling formula is as follows:

[0112] ;

[0113] wherein I sem (E j ) is the semantic importance index of the control E j , and the value range is [0, 1];

[0114] f role (E j ) is a control role level function, which defines the responsibility level of the control in the task flow (the numerical range is recommended to be set to 0~1, and the role is 1 for the main task button (“submit”, “confirm”, “next step”, etc.), and 0.2 for the auxiliary setting button (“language setting”, “theme color”, etc.), the control role level function f role (E j ) can be modeled by pre-preparing a recommended value rule table for each control;

[0115] f task (E j ) is a control task coverage function, which represents the degree of control coverage or support for the task target, and can also be modeled by pre-preparing a recommended value rule table for each control, for example, the scores of controls such as “medicine reminder”, “heart rate monitoring”, and “emergency contact person” are recommended to be 0.8~0.9, and the scores of controls such as “night mode switch” that are weakly associated with the task target are recommended to be 0.1~0.2;

[0116] f critical (E j ) is a control key function judgment function, which is a Boolean value function, and if E j belongs to a system-level core function, the value is 1, otherwise it is 0.

[0117] ω1 is the semantic importance index weight factor, ω2 is the control task coverage weight factor, and ω3 is the control key function judgment weight factor, and the constraint condition is ω1+ω2+ω3=1, and the initial values of each weight factor are set as: ω1=0.4, emphasizing the task responsibility position of the control, ω2=0.4, highlighting the coupling strength of the control and the task target, and ω3=0.2, as a Boolean enhancement item.

[0118] The semantic importance index modeling result I sem (E j ) will be used as the decision input of the “font size adjustment module” together with the interface element importance score S(E j ), forming a double-dimensional control mechanism, so as to ensure that low-frequency but high-semantic key controls can obtain appropriate font size display.

[0119] To this end, before the adjustment basis of the font size is obtained by using the standard deviation amplification mechanism, it is necessary to judge whether the control is a high-level semantic importance indicator, and for this purpose, a threshold I max When the semantic importance indicator modeling result I sem (E j ) is greater than the threshold I max , the system directly allocates the maximum font size to the control to ensure visibility, and optionally, I max It is recommended to set the range to 0.6-0.7 to ensure that only high semantic key controls enter the priority channel.

[0120] For example, in a user health management application, there are controls such as "medicine reminder", "heart rate monitoring", "emergency contact person", etc., which have low daily use frequency but are decisive in critical moments. In calculating the importance score of the interface element, the "heart rate monitoring" control has an interaction frequency of 1 time in the past 7 days, resulting in a lower S(E j ) score;

[0121] However, the semantic importance indicator modeling evaluates f role (E j ) of the control = 0.9 (main responsibility control), f task (E j ) = 0.85 (core task node), f critical (E j ) = 1 (system-level key), and the weight factors are ω1=0.4, ω2=0.4, ω3=0.2, respectively, then: I sem (E j ) The final calculation result is equal to 0.9, which is much higher than I max It is recommended to set the range to 0.6-0.7 (preferably 0.65), so the system directly allocates the maximum font size to the control.

[0122] Taking the "medicine reminder" control in the user health management application as an example, although its daily interaction frequency is not high, it has a direct impact on the user's health tasks. In traditional layout optimization, such controls are misjudged as secondary information due to low frequency, resulting in reduced font size and reduced recognition. The present application identifies the key attributes of the control through the semantic importance indicator and ensures that it displays the maximum font size in the interface, so that the user can quickly complete key information recognition and response operations under the premise of low visual cognitive load, significantly improving the overall interaction experience.

[0123] Further, the reordering and reconstruction of the page control uses a multi-objective optimization sorting algorithm based on stability constraints, taking the importance score of the interface element as the main sequence indicator, and combining historical layout inertia and user familiarity to construct the objective function:

[0124] ;

[0125] wherein, V(E j ) represents the target function score, F(E j ) represents the user familiarity, and the number of interactions N j and the residence time T j are weighted to obtain, and the formula is as follows:

[0126] ;

[0127] wherein, is the sum of the weighted product of the number of interactions and the residence time of all controls, that is, the standardization denominator, to ensure that each F(E j ) is normalized to the interval [0, 1];

[0128] D(E j ) represents the displacement distance of the current position information of the control from the last layout, and the formula is as follows:

[0129] ;

[0130] wherein, (x j (t) ,y j (t) ) represents the current coordinates of the control, and (x j (t-1) ,y j (t-1) ) represents the last layout coordinates;

[0131] η1, η2, and η3 are weight coefficients, η1 controls the importance of the current interaction, η2 controls the maintenance of the user familiarity to prevent cognitive burden caused by frequent rearrangement, and η3 controls the maintenance of the interface stability (minimum disturbance) to avoid the influence of large control migration on the interface stability.

[0132] Optionally, the recommended parameter settings are as follows (which can be adaptively adjusted):

[0133] η1∈[0.4, 0.6], which preferentially reflects the importance of the control;

[0134] η2∈[0.2, 0.4], which moderately retains user habits;

[0135] η3∈[0.1, 0.3], which appropriately reduces interface jumps;

[0136] The design basis of these coefficients is: the principle of priority of main functions + the principle of use memory protection + the principle of stable interaction experience.

[0137] The algorithm is based on a multi-objective optimization strategy with stability constraints, aiming to maximize user experience and interface convenience, by considering the importance score of the control, user familiarity, and layout inertia, to sort and optimize the interface controls, ensuring that important elements are displayed preferentially, while minimizing the user's adaptation cost to interface changes.

[0138] During operation, first, the S(E j ) corresponding to the control is collected j , the number of user interactions with the control, the average duration of stay, F(E j ) is calculated, and the displacement distance D(E j ) between the current position information and the position in the previous layout is calculated;

[0139] For each control E j , calculate its objective function score V(E j ), and then sort the controls according to the objective function score from high to low;

[0140] According to the sorting results, try to place the controls in the hot area, the secondary hot area, and the edge area in turn. If the pre-allocated position is occupied, start the local exchange search mechanism to try to exchange in a small range in the adjacent area to satisfy the local optimum.

[0141] If D(E j ) > D max (turbulence threshold, recommended D max = 30 pixels), then the original position is forced to remain unchanged to avoid frequent triggering of layout changes in a short period of time (cooling time can be set, such as only allowing rearrangement once in 60 seconds).

[0142] Further, the reordering and reconstruction of the page controls follow a joint optimization strategy of reachability and visual focus, with the following steps:

[0143] According to the importance score of the interface elements, all controls are divided into priority intervals, including high score zone (S(E j ) ≥ μ + σ), intermediate score zone (μ - σ < S(E j ) < μ + σ), and low score zone (S(E j ) ≤ μ - σ), where μ is the average value of S(E j ) of all controls, and σ is the standard deviation. This partitioning strategy is based on the discrete distribution characteristics of the standard normal distribution in statistics, ensuring that high-priority controls are significantly higher than the average, which helps to clarify the partition boundaries and facilitate interface mapping.

[0144] Distribute the controls in the high score zone in the hot area (such as the center of the interface, the bottom of the thumb reachable area);

[0145] The controls in the medium-score area are distributed in the secondary hot area (close to the edge of the hot area, can be quickly slid to reach the area);

[0146] The controls in the low-score area are concentrated in the edge area (less interactive area, such as top corner, hidden menu).

[0147] By prioritizing the scores of interface elements and placing high-priority elements in the user's hot zone, the convenience of using the interface and the accessibility of important functions can be improved.

[0148] Further, in the existing smart device interface design, due to physiological characteristics such as decreased vision and sensitive cognitive load, users often need larger fonts to ensure the readability of information. However, traditional interface font size design mostly uses fixed size or only based on overall scaling, and fails to combine the differences in user attention to each interface element in different interaction situations for personalized and dynamic adjustment, resulting in problems such as unclear information hierarchy, no differentiation between important information and secondary information, and redundant interface space occupation, reducing user interaction efficiency and comfort;

[0149] To solve the above technical problems, the present application provides a dynamic font size adjustment mechanism based on element importance score, which can ensure clear and readable key information while compressing the visual space occupied by non-key information, thereby realizing efficient allocation of visual resources and effectively improving the overall accessibility and readability of the interface.

[0150] The adjustment of the font size is based on the deviation of S(E j ) from the mean value μ of all interface element scores, and the amplification mechanism is implemented by standard deviation σ, and the specific formula is as follows:

[0151] ;

[0152] Wherein, μ represents the mean value of all interface element scores, as the reference font size, σ is the standard deviation, the larger the standard deviation, the greater the difference in element importance distribution, and the adjustment range of the font size also increases accordingly, θ is the amplification coefficient, which is used to control the amplification effect of the deviation on the font size to prevent excessive amplification or reduction. If the user group generally has severe vision loss, a larger value (such as 2.0~2.5) can be selected to significantly enhance the readability of the highlighted content. If the interface content hierarchy is not clear, a medium value (such as 1.5~2.0) can be selected. If the interface is already relatively simple, a smaller value (such as 1.0~1.5) can be selected to prevent excessive visual burden.

[0153] The deviation degree represents the relative importance of the element relative to the whole, and positive deviation corresponds to increasing font size, and negative deviation corresponds to decreasing font size.

[0154] The formula introduces normalized deviation amplification on the basis of the prior art, can linearly amplify or compress the font size according to the deviation degree of the element score relative to the overall mean value, realizes dynamic allocation of significant content font size increase and non-core content font size reduction, and the font size adjustment range is strictly controlled within the preset upper and lower limits min ,F max ]。

[0155] The core of the mechanism is to ensure that the important content font size is amplified, and the readability is improved, and the font size of unimportant elements is moderately compressed to release more visual space, and the purpose of the design is to solve the visual burden of the user, especially considering the problem of vision decline of many users, by dynamically adjusting the font size, the user is more comfortable and easy to understand in vision.

[0156] Further, in the traditional intelligent device interface system, interface adjustment (such as layout rearrangement, control position movement or size change) usually adopts instant refresh or fixed time length transition animation, which has two significant problems: the user's perception sensitivity to interface dynamic change is high, if the animation time length cannot be flexibly adapted according to the current interface adjustment complexity and historical operation load, it may cause distraction, anxiety, and understanding delay; and the existing method is a static animation scheme (such as setting a uniform 300ms time length), which cannot comprehensively consider the interface change amplitude (control displacement amount, size adjustment amplitude) and the user's familiarity and cognitive adaptation ability to the interface dynamics in the past;

[0157] In view of the above problems, the present application provides a cognitive load driven dynamic animation transition mechanism, which can adjust the animation time length and transition smoothness on demand according to the user's historical operation load and the current interface change amplitude, to maximize the friendliness and naturalness of interface adjustment;

[0158] The present application realizes the interface animation transition mechanism through the following core steps:

[0159] Firstly, collect the user's historical cognitive load data: including the average stay time (reflecting the understanding speed), the operation error rate (reflecting the acceptance degree) and the stay time of the last interface rearrangement;

[0160] Secondly, calculate the total visual displacement caused by this interface rearrangement: sum the displacement distances of all the adjusted controls to quantitatively reflect the change amplitude;

[0161] Thirdly, the animation transition mechanism is driven by a cognitive load model, and the animation time T anim is determined according to the following formula:

[0162] ;

[0163] Wherein, T0 is the basic shortest animation duration (optionally, T0 = 300~500ms), to avoid too short to cause visual jump, C hist represents the user's average cognitive load, represents the user's understanding pressure in long-term interaction, defined as the normalized combination of the average interface dwell time and the operation error rate of the user in a number of historical operations, C δ represents the total visual displacement of the control caused by this rearrangement, the unit can be defined as pixel or normalized relative displacement, the weighted average of the two is taken as the median, further enhancing the smoothness of extreme values (such as an abnormal high), ρ is the response coefficient, used to control the amplification of the change range and the cognitive load on the animation duration, to ensure that the animation is slow and natural when the cognitive load is high, the recommended value: ρ = 500~800ms, if the load is moderate (C hist +C δ ≈1), the animation increase is about 250~400ms, and the increase is more than 800ms when the load is high.

[0164] Specifically, the user's average cognitive load C hist is calculated as follows:

[0165] ;

[0166] Wherein, represents the historical average dwell time (seconds), T ref is the reference time (recommended 10s), E rate is the historical error rate (0~1), λ err is the error rate weighting coefficient, preferably, λ err is 0.5.

[0167] The total visual displacement of the control caused by this rearrangement C δ is calculated as follows:

[0168] ;

[0169] If the historical load is high (C hist →1), but the current change is small (C δ →0), the animation still needs to be lengthened to maintain the sense of cognitive safety, if the current change is severe and the historical adaptability is good, the animation duration still rises with the change, both of which are high, the animation slows down significantly;

[0170] Finally, the animation transition based on gradual interpolation is executed, and through non-linear interpolation (such as ease-in-out curve) in the entire animation execution period, visual smoothness and reduction of abruptness are ensured.

[0171] The animation transition mechanism is driven by a cognitive load model. It achieves a smooth transition of interface changes by dynamically adjusting the duration of the animation. The higher the user's cognitive load, the slower the interface change should be to reduce visual fatigue caused by interface adjustments. This mechanism aims to improve the user's interface experience and avoid abrupt interface changes from causing inconvenience to the user.

[0172] For example, in the interface adjustment process of this invention, the animation transition mechanism performs the following operations in sequence:

[0173] Collect historical data: The system analyzes the average dwell time and error rate of the user's last 20 operations;

[0174] Analyze this layout adjustment: Calculate the displacement of all controls and generate C. δ ;

[0175] Calculate the animation duration using the formula:

[0176] Substitute:

[0177] T0=400ms, ρ=600ms

[0178] Assumption:

[0179] C hist =0.8,C δ =0.6

[0180] but: ;

[0181] Set animation parameters: Animation interpolation function: ease-in-out (smooth curve), animation duration: 820ms;

[0182] Perform control animation: The position and size of the control change gradually, and the interface prompts and voice broadcasts are triggered simultaneously.

[0183] Furthermore, the interface layout recommendation method also includes the following steps:

[0184] After the interface adjustment is completed, voice prompts are automatically generated to point out the changed areas or controls in the interface through voice broadcast. The prompts include the control name and the type of change.

[0185] The voice prompt information is generated using a Top-K change region selection strategy, specifically:

[0186] The degree of change of all controls in the calculation interface, including weighted scores for changes in position, size, color, or function;

[0187] Sort and select the n controls with the highest degree of change;

[0188] Generate voice template prompts for each control.

[0189] The generation scheme of the voice prompt adopts a Top-K change area selection strategy, helps the user to more easily understand the changes in the interface, calculates the degree of control change and generates a voice prompt for the control with larger changes, and the user can quickly know which interface elements have changed, which is particularly important for the user because they may have difficulty quickly identifying the interface changes through vision.

[0190] It should be noted that although the existing smart devices are relatively popular, the smart devices used by the user may sometimes be low-end models with weak processing performance and limited memory capacity (typical configuration: less than 3GB of memory, CPU frequency not higher than 1.5GHz), in order to balance the intelligence of interface layout reconstruction and the stability of system operation, the method presets two schemes of standard UI template and lightweight UI template in the system architecture, and judges whether the current running environment is a low-end model, if the user uses a low-end model with weak processing performance and limited memory capacity, then:

[0191] The lightweight UI template reduces CPU and memory load by reducing animation elements, compressing image resources, and simplifying control styles;

[0192] The low-end model loads the lightweight UI template by default when starting, and only retains the core control element reordering and font size adjustment logic;

[0193] The score S(E j ) of the interface element and the layout optimization sorting process can still be performed in the low-resource mode, but the result is embodied as low-frequency reconstruction in visual presentation to avoid frequent changes;

[0194] And when the running environment is the lightweight UI template, a static switching mode is used for presentation.

[0195] By setting the two schemes of standard UI template and lightweight UI template, it is ensured that the layout optimization method is still executable in the case of limited device resources, does not affect the overall stable operation of the system, and is more suitable for the user's daily use in cooperation with the stable preference control mechanism.

[0196] Finally, it should be noted that: the above only describes the preferred embodiments of the present application and is not intended to limit the present application, although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent replacements for some technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for recommending interface layouts for smart devices, characterized in that, Includes the following steps: Acquire user behavior data, which includes historical behavior data and real-time operation data. The user behavior data includes the user's click operation coordinates, the type of interaction performed, the duration of each interaction, the type of operation intent of the current task, and the result identifier of the interaction operation. Build a user behavior model based on the user behavior data. The user behavior data is mapped into structured behavior vectors, and the operation behaviors at different time points are integrated. The behavior weight of each behavior record is adjusted according to the time decay factor to construct a user time-series behavior model that reflects the user's current usage habits and preferences. Based on the multidimensional data in the user behavior data, a weighted integral algorithm is used to calculate the importance score of each interface element, and the weight of each dimension is adjusted based on the user profile update strategy. The importance score of the interface element is determined by the following formula: ； Among them, E j Represented as a UI element, S(E) j f represents the importance score of the interface element. click (E j f represents the click frequency of this element. stay (E j f represents the average time a user spends on this element. path (E j f represents the frequency of the element in the user's access path. success (E j f is the success rate of the operation on this element. intent (E j ) represents the intensity index of the operation related to the current interaction intent, where α, β, γ, δ, and ε are the weighting coefficients, respectively; Based on the importance scores of the interface elements, the page controls are reordered and reconstructed, with high-scoring interface elements given priority in the main visual area, and controls with the same score are further sorted according to the historical hot zone priority. The reordering and reconstruction of page controls adopts a multi-objective optimization sorting algorithm based on stability constraints. The algorithm uses the importance score of interface elements as the main order index and combines historical layout inertia and user familiarity to construct the objective function. ； Among them, V(E) j F(E) represents the objective function score. j The value D(E) represents user familiarity, calculated by weighting the number of times users interact with the control and the duration of their interaction. j ) represents the current position information of the control and the displacement distance in the previous layout, and η1, η2, and η3 are the weight coefficients; Based on users' historical font size usage behavior, adjust the font size of each interface element, increasing the font size of interface elements with high importance scores and decreasing the font size of interface elements with low importance scores; During each reordering and reconstruction of the interface, an animation transition mechanism driven by the current change magnitude and the user's historical cognitive load is adopted to achieve a smooth transition in the interface change process. The animation transition mechanism is driven by a cognitive load model, with an animation duration T. anim It is determined by the following formula: ； Among them, T0 is the shortest animation duration based on the base, C hist C represents the user's historical average cognitive load. δ This represents the total visual displacement of the controls caused by this rearrangement, where ρ is the response coefficient.

2. The interface layout recommendation method for smart devices according to claim 1, characterized in that, The user behavior model is constructed as follows: ； Where n represents the number of user operation records; w i The time decay weight for the i-th behavior record is represented as: ； Where λ is a parameter controlling the degree of attenuation of the influence of historical behavior, and T now T is the current system timestamp. i The timestamp of the i-th action record; u i The feature vector representing the i-th operation is denoted as: ； Where, p i The click position is represented by two-dimensional coordinates, t i For operation type, d i a represents the duration the element remains in the room. i For the category of operational intent, r i This is the result of the operation.

3. The interface layout recommendation method for smart devices according to claim 2, characterized in that, The reordering and reconstruction of page controls follows a joint optimization strategy of accessibility and visual focus, and the steps are as follows: All controls are divided into priority zones based on the importance score of the interface elements, including high score zone, medium score zone and low score zone. Distribute controls in high-scoring areas to hot zones; Distribute the controls for the middle scoring area to the secondary hot zone; Concentrate controls in low-scoring areas at the edge.

4. The interface layout recommendation method for smart devices according to claim 2, characterized in that, The font size adjustment is based on S(E) j The deviation from the mean μ of the scores of all interface elements is amplified by the standard deviation σ, as shown in the following formula: ； Where μ represents the mean score of all interface elements, σ ​​is the standard deviation, and θ is the amplification factor.

5. The interface layout recommendation method for smart devices according to claim 1, characterized in that, The interface layout recommendation method also includes the following steps: After the interface adjustment is completed, voice prompts are automatically generated to indicate the changed areas or controls on the interface through voice broadcast. The prompts include the control name and the type of change.

6. The interface layout recommendation method for smart devices according to claim 5, characterized in that, The voice prompt information is generated using a Top-K change region selection strategy, specifically: Calculate the degree of change of all controls in the calculation interface; Sort and select the n controls with the highest degree of change; Generate voice template prompts for each control.

Images