Intelligent recommendation system for smart parks based on data analysis

Abstract

The present invention relates to the field of intelligent recommendation technology, specifically to a smart park intelligent recommendation system based on data analysis, which includes a behavior pattern analysis module, a user preference dynamic update module, an activity recommendation adjustment module, and a feedback processing module. In the present invention, by real-time tracking of user activity patterns in the park and dynamically updating user portraits, the level of service personalization is significantly improved, breaking through the limitations of traditional recommendation systems, and by comparing historical and current data, adjusting preference weights, so that the recommended content more accurately reflects the user's immediate needs, and automatically adjusting the recommendation list and synchronously updating the activity database at the same time, so that the recommendation service is not only responsive, but also highly synchronized with the user's actual behavior and change trends. The continuous optimization and update mechanism ensures the high relevance and personalization of the recommended content, greatly improving the user's interactive experience and satisfaction.

Description

Technical Field

[0001] The present invention relates to the field of intelligent recommendation technology, and in particular to a smart park intelligent recommendation system based on data analysis. Background Art

[0002] The field of intelligent recommendation technology usually relies on machine learning and artificial intelligence algorithms, such as collaborative filtering, content recommendation and deep learning, to analyze large amounts of user behavior data and preference settings to provide targeted product, service or content recommendations. The application range is wide, including e-commerce, online media, social networking platforms, and more specific environments such as smart parks. Through these technologies, the system can automatically improve the user experience and increase the relevance and attractiveness of services or products.

[0003] Among them, the smart park intelligent recommendation system mainly uses intelligent recommendation technology to optimize the service and facility utilization efficiency within the park, and provides customized suggestions by analyzing the behavior patterns and preferences of park users, such as activity recommendations, dining options, transportation methods, etc. The main purpose of the system is to improve user satisfaction and park operation efficiency, while providing data support for park managers to carry out more effective resource allocation and service optimization.

[0004] Existing intelligent recommendation systems for smart parks have obvious deficiencies in adaptability and real-time responsiveness in dynamic environments. Although they can process large amounts of static data and provide recommendations based on past behavior, they are generally unable to effectively capture immediate changes in user behavior. Algorithms that lack flexibility make it difficult to update user portraits and recommendation lists in real time, resulting in service recommendations that fail to reflect users' latest preferences and needs in a timely manner. In addition, when user preferences change rapidly, delayed updates in traditional systems may cause recommendations to lose relevance, reducing user engagement and satisfaction. Lag is particularly prominent in a smart park environment where demand changes rapidly, affecting the park's service efficiency and user experience. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a smart park intelligent recommendation system based on data analysis.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions: The smart park intelligent recommendation system based on data analysis includes:

[0007] The behavior pattern analysis module collects user movement and stay time points based on the monitoring data in the smart park, analyzes user time series data to identify user activity peaks, adjusts the collection frequency to match the user activity cycle, statistically analyzes user behavior patterns, and generates user behavior pattern data;

[0008] The user preference dynamic update module calculates the user's stay time in different areas through the user behavior pattern data, compares historical data to mine user behavior trends, adjusts preference weights, and generates dynamically updated user portraits;

[0009] The activity recommendation adjustment module uses the dynamically updated user portrait to evaluate the user's response to the current activity, collect the user's interactive data on the activity, synchronously update the activity database to reflect the user's latest preferences, and generate an activity recommendation list;

[0010] The feedback processing module analyzes user satisfaction based on the activity recommendation list and user feedback on park facilities, adjusts facility recommendations according to satisfaction data, dynamically updates the activity recommendation list to match user expectations, and generates an intelligent recommendation plan.

[0011] As a further solution of the present invention, the activity peak identification step is:

[0012] Based on the monitoring data in the smart park, collect user movement and residence time data, count the number of users at each time point, and obtain the user density time series;

[0013] Apply the moving average algorithm to the user density time series to generate a smoothed user density series;

[0014] Detecting activity peaks based on the smoothed user density sequence, ;

[0015] in, is the activity peak detection data, It's time point The user density represents the number of users at each time point. is the average user density, reflecting the normal level of the overall user density. It's time point The weights are adjusted according to the importance of the different time periods to optimize the calculation flexibility. is the total number of time points;

[0016] Significant peak time points are extracted from the activity peak detection data and marked to reveal the intensive periods of user activities, and the activity peak recognition results are output.

[0017] As a further solution of the present invention, the steps of acquiring the user behavior pattern data are:

[0018] Call the activity peak recognition result, calculate the activity period estimation result, analyze the interval of the peak time period, obtain the activity period estimation data set, and calculate the new collection frequency. ;

[0019] in, is the new acquisition frequency, is the average activity period during peak hours, calculated as the average of the intervals during all peak hours, and are the shortest and longest time intervals among all activity cycles, is the adjustment coefficient, which is used to optimize the acquisition frequency and obtain the adjusted acquisition frequency data set;

[0020] Calling the adjusted collection frequency data set to collect user activity data to obtain an updated user activity data set;

[0021] Using the updated user activity data set, calculate the user's stay frequency and movement path in different time periods, and conduct preliminary behavior pattern analysis. ;

[0022] in, is the weighted average behavior index, Indicates Behavioral data of each user, including dwell time and movement frequency, is the behavior weight factor, which is allocated according to the importance of user behavior. is the total number of users, and we get a preliminary set of user behavior pattern analysis results;

[0023] Through the preliminary user behavior pattern analysis result set, weighted and normalized processing is performed, ;

[0024] in, It is The number of occurrences of the class behavior pattern, is the weight, determined by the importance or frequency of user behavior, is the smoothing parameter used to adjust the effect of the standard deviation, is the standard deviation of the behavior pattern, It is the total category of behavior patterns, and obtains user behavior pattern data .

[0025] As a further solution of the present invention, the calculation steps of the residence time data of the difference area are:

[0026] Through user behavior pattern data, capture the user's movement and interaction behavior in the park, record timestamps and geographic location tags, and generate a preliminary behavior data set;

[0027] From the preliminary behavior data set, all activity records of the user in the target area are filtered using the geographic location tag to obtain a partitioned behavior data set;

[0028] Calculate the residence time. ;

[0029] in, Indicates that the user is in the target area The total residence time, is the target area, is the number of events recorded in the target region, is the total activity time, It is The duration of the activity, Yes Activity The impact factor adjusts the weight of each activity. is an adjustment parameter used to stabilize the small amount of the divisor and avoid the denominator being zero. It is an adjustment parameter used to adjust the nonlinear intensity of the calculation, affect the sensitivity of the final residence time, and calculate the user's residence time data in the difference area.

[0030] As a further solution of the present invention, the step of obtaining the dynamically updated user portrait is:

[0031] Retrieving the residence time data of the difference area and merging it with the stored historical residence time data to generate a comprehensive data set;

[0032] Performing time series analysis on the comprehensive data set to identify significant patterns in the data and obtain a time series analysis result set;

[0033] According to the time series analysis result set, trend identification is performed. ;

[0034] in, Indicates that in the target area The trend strength of user behavior within Indicates the target area, is the total number of time points, It is in Time point in target area User stay time, yes The average value of yes The rate of change at a time point is expressed as and The difference, is the normalization constant used The impact of adjustment ensures that comparisons between different regions are comparable, express The average value of , generates the user behavior trend analysis results;

[0035] According to the user behavior trend analysis results, the weights are adjusted to obtain a dynamically updated user portrait.

[0036] As a further solution of the present invention, the step of evaluating the user's response to the current activity is:

[0037] Extracting the user's behavior log and preference settings from the dynamically updated user portrait to obtain user behavior and preference data;

[0038] Based on the user behavior and preference data, calculate the user's responsiveness to the activity, ;

[0039] in, is the user's responsiveness to the target activity, Indicates the user's The intensity of preference for class activities, Indicates The attractiveness of similar activities is analyzed based on the content of the activities and historical user engagement. is the exponential weight of the activity's attractiveness, enhancing the impact of the activity's attractiveness in the responsiveness calculation, Is the user and activity type The distance or mismatch between is the attenuation factor, used to adjust for the effect of distance or mismatch, Indicates the number of activities;

[0040] The calculated Combined with user portraits, the user's overall reaction to the activity is evaluated to obtain comprehensive response evaluation data.

[0041] As a further solution of the present invention, the steps of obtaining the activity recommendation list are:

[0042] Collect user interaction data on activities through the user interface and calculate the user preference weight for each activity:

[0043] ;

[0044] in, is the updated value of the user preference, is the user's feedback rating for the nth activity, is the frequency of interaction, indicating the popularity of the activity. is the number of days since the event was launched, used to adjust for the effect of preferences changing over time. is a constant, get the user preference update data, Indicates the number of activities;

[0045] According to the comprehensive response evaluation data and user preference update data, each activity recommendation score is recalculated and the activity list is sorted, ;

[0046] in, is the target activity recommendation score, Is the user's Response score for each activity, is the updated activity weight, Indicates the number of activities and generates optimized recommendation list data;

[0047] Synchronize the optimized recommendation list data to the user interface to ensure that activities with high scores are displayed first in the activity recommendation list.

[0048] As a further solution of the present invention, the steps of obtaining the intelligent recommendation solution are:

[0049] Collect user feedback on park facilities, obtain original data on facility satisfaction, and calculate facility satisfaction scores. ;

[0050] in, is the satisfaction score of the facility, It is User ratings for facilities is the time since the rating, is the time decay coefficient, which is used to adjust the weight of old ratings, is a constant, Generate comprehensive satisfaction analysis data for the total number of facilities;

[0051] Evaluate the performance of each facility based on the comprehensive satisfaction analysis data, determine whether adjustments are needed based on the user expectation match, and calculate the adjustment recommendation priority. ;

[0052] in, is the recommended priority of the target facility, is the satisfaction score of the facility, is the average of the satisfaction scores, is the standard deviation of satisfaction scores;

[0053] according to , dynamically update the activity recommendation list to match user expectations and obtain an intelligent recommendation solution.

[0054] Compared with the prior art, the advantages and positive effects of the present invention are:

[0055] In the present invention, by real-time tracking of user activity patterns in the park and dynamically updating user portraits, the level of service personalization is significantly improved, breaking through the limitations of traditional recommendation systems. By comparing historical and current data and adjusting preference weights, the recommended content can more accurately reflect the user's immediate needs. At the same time, the recommendation list is automatically adjusted and the activity database is synchronously updated, so that the recommendation service is not only timely, but also highly synchronized with the user's actual behavior and change trends. The continuous optimization and update mechanism ensures the high relevance and personalization of the recommended content, greatly improving the user's interactive experience and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a system flow chart of the present invention;

[0057] Figure 2 It is a flow chart for identifying the activity peak of the present invention;

[0058] Figure 3 A flowchart for obtaining user behavior pattern data of the present invention;

[0059] Figure 4 A flow chart for calculating the residence time data of the difference area of ​​the present invention;

[0060] Figure 5 A flowchart for obtaining a dynamically updated user portrait of the present invention;

[0061] Figure 6 A flow chart for evaluating the user's response to the current activity of the present invention;

[0062] Figure 7 A flowchart for obtaining an activity recommendation list of the present invention;

[0063] Figure 8 This is a flow chart for obtaining the intelligent recommendation solution of the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0065] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0066] Example 1: Please refer to Figure 1 , the smart park intelligent recommendation system based on data analysis includes:

[0067] The behavior pattern analysis module collects user movement and stay time points based on the monitoring data in the smart park, analyzes user time series data to identify user activity peaks, adjusts the collection frequency to match the user activity cycle, statistically analyzes user behavior patterns, and generates user behavior pattern data;

[0068] The user preference dynamic update module calculates the user's stay time in different areas through user behavior pattern data, compares historical data to mine user behavior trends, adjusts preference weights, and generates dynamically updated user portraits;

[0069] The activity recommendation adjustment module uses dynamically updated user portraits to evaluate users' responses to current activities, collects user interaction data on activities, synchronously updates the activity database to reflect users' latest preferences, and generates an activity recommendation list;

[0070] The feedback processing module analyzes user satisfaction based on the activity recommendation list and combined with user feedback on park facilities, adjusts facility recommendations based on satisfaction data, dynamically updates the activity recommendation list to match user expectations, and generates intelligent recommendation plans.

[0071] User behavior pattern data specifically include user activity frequency collection results, time point analysis results and user behavior cycle. The dynamically updated user portrait includes preference change tracking results, portrait update frequency and behavior preference analysis results. The activity recommendation list includes activity matching index and recommended activity types. The intelligent recommendation plan includes facility feedback integration results, satisfaction survey results and recommendation adaptability evaluation results.

[0072] See also Figure 2 , the steps for identifying the activity peak are:

[0073] Based on the monitoring data in the smart park, collect user movement and residence time data, count the number of users at each time point, and obtain the user density time series;

[0074] Apply the moving average algorithm to the user density time series to generate a smoothed user density series;

[0075] According to the smoothed user density sequence, the activity peak is detected. ;

[0076] in, is the activity peak detection data, It's time point The user density represents the number of users at each time point. is the average user density, reflecting the normal level of the overall user density. It's time The weights are adjusted according to the importance of the different time periods to optimize the calculation flexibility. is the total number of time points;

[0077] Significant peak time points are extracted from activity peak detection data and marked to reveal the intensive periods of user activities and output activity peak identification results.

[0078] Collect data:

[0079] First, collect each time point from the monitoring system The number of users data generates a time series data set .

[0080] Calculate the average user density :

[0081] ;

[0082] Assume the following user density data:

[0083] ;

[0084] ;

[0085] ;

[0086] ;

[0087] but, ;

[0088] Set different weights for different time points as needed ,set up:

[0089] ;

[0090] ;

[0091] ;

[0092] Compute the weighted squared difference:

[0093] ;

[0094] ;

[0095] ;

[0096] ;

[0097] ;

[0098] but, ;

[0099] This effectively identifies activity peaks in time series.

[0100] See also Figure 3 , the steps to obtain user behavior pattern data are:

[0101] Call the activity peak recognition results, calculate the activity period estimation results, analyze the intervals of the peak time periods, obtain the activity period estimation data set, and calculate the new collection frequency. ;

[0102] in, is the new acquisition frequency, is the average activity period during peak hours, calculated as the average of the intervals during all peak hours, and are the shortest and longest time intervals among all activity cycles, is the adjustment coefficient, which is used to optimize the acquisition frequency and obtain the adjusted acquisition frequency data set;

[0103] Calling the adjusted collection frequency data set to collect user activity data to obtain an updated user activity data set;

[0104] Using the updated user activity data set, calculate the user's stay frequency and movement path in different time periods, and conduct preliminary behavior pattern analysis. ;

[0105] in, is the weighted average behavior index, Indicates Behavioral data of each user, including dwell time and movement frequency, is the behavior weight factor, which is allocated according to the importance of user behavior. is the total number of users, and we get a preliminary set of user behavior pattern analysis results;

[0106] Through the preliminary user behavior pattern analysis result set, weighted and normalized processing is performed. ;

[0107] in, It is The number of occurrences of the class behavior pattern, is the weight, determined by the importance or frequency of user behavior, is the smoothing parameter used to adjust the effect of the standard deviation, is the standard deviation of the behavior pattern, It is the total category of behavior patterns, and obtains user behavior pattern data .

[0108] Calculate and adjust the acquisition frequency:

[0109] Hypothetical Activity Cycle Dataset For 10, 15, 20.12, 18, set Set to 0.5 to adjust frequency sensitivity.

[0110] Calculate the average activity period :

[0111] ;

[0112] Find the minimum and maximum activity periods and :

[0113] ;

[0114] Calculate the adjusted acquisition frequency :

[0115] ;

[0116] Preliminary analysis of user behavior patterns:

[0117] Assume there are 5 users’ behavior data (representing a certain activity index) and setting weight factors for these behavioral data .

[0118] Compute the sum of weighted behavioral data:

[0119] ;

[0120] calculate :

[0121] ;

[0122] result It can be interpreted as a weighted average behavior index, reflecting the relative importance of different user behaviors.

[0123] Step 3 output This provides a behavioral weighted average perspective for the analysis in step 4, which is then further refined to form a more detailed and dynamically adjusted behavioral pattern data by considering the variability and relative importance of each behavioral pattern. The calculation in step 4 does not directly use the results of step 3, but is based on the same data set. and behavioral markers , and add weight and standard deviation calculations to output more detailed and applicable user behavior pattern data.

[0124] User behavior pattern data generation:

[0125] Assume there are three behavior modes ,in , weight [0.2, 0.5, 0.3], standard deviation ,set up .

[0126] Calculate the weighted impact of the total number and standard deviation of behavior patterns:

[0127] ;

[0128] Calculate the contribution of each behavior mode :

[0129] for :

[0130] ;

[0131] for :

[0132] ;

[0133] for :

[0134] ;

[0135] To summarize all contributions:

[0136] ;

[0137] Output generated user behavior pattern data .

[0138] See also Figure 4 , the calculation steps of the residence time data of the difference area are:

[0139] Through user behavior pattern data, capture the user's movement and interaction behavior in the park, record timestamps and geographic location tags, and generate a preliminary behavior data set;

[0140] From the preliminary behavior dataset, all activity records of users in the target area are filtered using geographic location tags to obtain a partitioned behavior dataset.

[0141] Calculate the residence time. ;

[0142] in, Indicates that the user is in the target area The total residence time, is the target area, is the number of events recorded in the target region, is the total activity time, It is The duration of the activity, Yes Activity The impact factor adjusts the weight of each activity. is an adjustment parameter used to stabilize the small amount of the divisor and avoid the denominator being zero. It is an adjustment parameter used to adjust the nonlinear intensity of the calculation, affect the sensitivity of the final residence time, and calculate the user's residence time data in the difference area.

[0143] Assume that in a specific area , we have the following activity records:

[0144] Number of activities ;

[0145] Duration minute;

[0146] Also set:

[0147] Activity Impact Factor ;

[0148] parameter ;

[0149] parameter .

[0150] Calculate total active time:

[0151] ;

[0152] Calculate the weighted time for each activity:

[0153] First activity:

[0154] ;

[0155] Second activity:

[0156] ;

[0157] The third activity:

[0158] ;

[0159] Calculate total dwell time :

[0160] ;

[0161] Shows the user in the area The total residence time is .

[0162] See also Figure 5 , the steps to obtain the dynamically updated user portrait are:

[0163] Recalling the residence time data of the difference area and merging it with the stored historical residence time data to generate a comprehensive data set;

[0164] Perform time series analysis on the comprehensive data set, identify significant patterns in the data, and obtain a time series analysis result set;

[0165] Based on the time series analysis result set, trend identification is performed. ;

[0166] in, Indicates that in the target area The trend strength of user behavior within Indicates the target area, is the total number of time points, It is in Time point in target area User stay time, yes The average value of yes The rate of change at a time point is expressed as and The difference, is the normalization constant used The impact of adjustment ensures that comparisons between different regions are comparable, express The average value of , generates the user behavior trend analysis results;

[0167] According to the results of user behavior trend analysis, the weights are adjusted to obtain a dynamically updated user portrait.

[0168] Assume that in the area Observation time point , specific residence time data 30 minutes, 45 minutes and 40 minutes respectively.

[0169] calculate :

[0170] ;

[0171] Calculate the time points (Rate of Change):

[0172] (There is no previous time point, set to 0);

[0173] ;

[0174] ;

[0175] calculate :

[0176] ;

[0177] set up , used to adjust impact.

[0178] calculate :

[0179] First calculate the contribution of each time point:

[0180] ;

[0181] ;

[0182] ;

[0183] Add up and take the square root:

[0184] ;

[0185] pass The introduction of , emphasizes the dynamic importance of changes in time series data, and and The use of provides a standardized treatment of the rate of change, ensuring the stability and comparability of the calculation results.

[0186] See also Figure 6 , the evaluation steps of the user's response to the current activity are:

[0187] Extract user behavior logs and preference settings from dynamically updated user portraits to obtain user behavior and preference data;

[0188] Calculate user responsiveness to activities based on user behavior and preference data, ;

[0189] in, is the user's responsiveness to the target activity, Indicates the user's The intensity of preference for class activities, Indicates The attractiveness of similar activities is analyzed based on the content of the activities and historical user engagement. is the exponential weight of the activity's attractiveness, enhancing the impact of the activity's attractiveness in the responsiveness calculation, Is the user and activity type The distance or mismatch between is the attenuation factor, used to adjust for the effect of distance or mismatch, Indicates the number of activities;

[0190] The calculated Combined with user portraits, the user's overall reaction to the activity is evaluated to obtain comprehensive response evaluation data.

[0191] Assume there are three types of activities, and the user's preference intensity for the activities They are 0.8, 0.5, and 0.3 respectively. They are 0.9, 0.7 and 0.4 respectively.

[0192] set up and , the mismatch between users and activities They are 1, 2, and 3 respectively.

[0193] Calculate the weighted index of attractiveness for each activity:

[0194] ;

[0195] ;

[0196] ;

[0197] Calculate the decay factor for each activity:

[0198] ;

[0199] ;

[0200] ;

[0201] Calculate the responsivity:

[0202] ;

[0203] ;

[0204] ;

[0205] ;

[0206] Derive an evaluation of the user's response to the activity.

[0207] See also Figure 7 , the steps to obtain the activity recommendation list are:

[0208] Collect user interaction data on activities through the user interface and calculate the user preference weight for each activity:

[0209] ;

[0210] in, is the updated value of the user preference, is the user's feedback rating for the nth activity, is the frequency of interaction, indicating the popularity of the activity. is the number of days since the event was launched, used to adjust for the effect of preferences changing over time. is a constant, get the user preference update data, Indicates the number of activities;

[0211] Update the data based on the comprehensive response evaluation data and user preferences, recalculate the recommendation score of each activity, and sort the activity list, ;

[0212] in, is the target activity recommendation score, Is the user's Response score for each activity, is the updated activity weight, generating optimized recommendation list data, Indicates the number of activities;

[0213] Synchronize the optimized recommendation list data to the user interface to ensure that activities with high scores are displayed first in the activity recommendation list.

[0214] Assume there are three activities, user feedback and activity information are as follows:

[0215] Activity 1: ;

[0216] Activity 2: ;

[0217] Activity 3: ;

[0218] Calculate the freshness impact factor ( =2.71):

[0219] For Activity 1: ;

[0220] For Activity 2: ;

[0221] For Activity 3: ;

[0222] Calculate the updated value of user preference:

[0223] ;

[0224] ;

[0225] but, ;

[0226] Recommendation score calculation:

[0227] Assume that the updated activity weights and user responses are as follows:

[0228] Activity 1: ;

[0229] Activity 2: ;

[0230] Activity 3: ;

[0231] Calculate the recommendation score:

[0232] ;

[0233] ;

[0234] ;

[0235] but, ;

[0236] Combination and , can quickly understand and adapt to user preferences by dynamically adjusting recommendation lists to match users’ current interests and behaviors.

[0237] See also Figure 8 , the steps to obtain the intelligent recommendation solution are:

[0238] Collect user feedback on park facilities, obtain original data on facility satisfaction, and calculate facility satisfaction scores. ;

[0239] in, is the satisfaction score of the facility, It is User ratings for facilities is the time (days) since the rating, is the time decay coefficient, which is used to adjust the weight of old ratings, is a constant, Generate comprehensive satisfaction analysis data for the total number of facilities;

[0240] Evaluate the performance of each facility based on comprehensive satisfaction analysis data, determine whether adjustments are needed based on the degree of match with user expectations, and calculate the priority of adjustment recommendations. ;

[0241] in, is the recommended priority of the target facility, is the satisfaction score of the facility, is the average of the satisfaction scores, is the standard deviation of satisfaction scores;

[0242] according to , dynamically update the activity recommendation list to match user expectations and obtain intelligent recommendation solutions.

[0243] Assume that there are three facilities with the following rating data:

[0244] Facility 1: sky;

[0245] Facility 2: sky;

[0246] Facility 3: sky;

[0247] set up .

[0248] Calculate the decay weight for each facility ( =2.71):

[0249] Facility 1: ;

[0250] Facility 2: ;

[0251] Facility 3: ;

[0252] Calculate the numerator and denominator of the satisfaction score:

[0253] ;

[0254] ;

[0255] but, ;

[0256] Calculate the mean satisfaction and standard deviation:

[0257] ;

[0258] Compute the squared deviation for each score:

[0259] Facility 1:

[0260] ;

[0261] Facility 2:

[0262] ;

[0263] Facility 3:

[0264] ;

[0265] Sum and calculate standard deviation:

[0266] The standard deviation is obtained by summing the squares of the above deviations, dividing by the number of facilities (here 3), and taking the square root:

[0267] ;

[0268] ;

[0269] ;

[0270] Recommended priorities for computing facilities:

[0271] Substitute values ​​into the formula:

[0272] Facility 1:

[0273] ;

[0274] Facility 2:

[0275] ;

[0276] Facility 3:

[0277] ;

[0278] Then we can get the recommendation priority: Facility 1>Facility 2>Facility 3.

[0279] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. The intelligent recommendation system for smart parks based on data analysis is characterized by: The smart park intelligent recommendation system based on data analysis includes: The behavior pattern analysis module collects user movement and stay time points based on the monitoring data in the smart park, analyzes user time series data to identify user activity peaks, adjusts the collection frequency to match the user activity cycle, statistically analyzes user behavior patterns, and generates user behavior pattern data; The user preference dynamic update module calculates the user's stay time in different areas through the user behavior pattern data, compares historical data to mine user behavior trends, adjusts preference weights, and generates dynamically updated user portraits; The activity recommendation adjustment module uses the dynamically updated user portrait to evaluate the user's response to the current activity, collect the user's interactive data on the activity, synchronously update the activity database to reflect the user's latest preferences, and generate an activity recommendation list; The feedback processing module analyzes user satisfaction based on the activity recommendation list and user feedback on park facilities, adjusts facility recommendations based on satisfaction data, dynamically updates the activity recommendation list to match user expectations, and generates an intelligent recommendation plan; The calculation steps of the residence time data of the difference area are: Through user behavior pattern data, capture the user's movement and interaction behavior in the park, record timestamps and geographic location tags, and generate a preliminary behavior data set; From the preliminary behavior data set, all activity records of the user in the target area are filtered using the geographic location tag to obtain a partitioned behavior data set; Calculate the residence time. ; in, Indicates that the user is in the target area The total residence time, is the target area, is the number of events recorded in the target region, is the total activity time, It is The duration of the activity, Yes Activity The influence factor of each activity is adjusted. is an adjustment parameter used to stabilize the small amount of the divisor and avoid the denominator being zero. It is an adjustment parameter used to adjust the nonlinear intensity of the calculation, affect the sensitivity of the final residence time, and calculate the user's residence time data in the difference area; The steps for obtaining the dynamically updated user portrait are as follows: Retrieving the residence time data of the difference area and merging it with the stored historical residence time data to generate a comprehensive data set; Performing time series analysis on the comprehensive data set to identify significant patterns in the data and obtain a time series analysis result set; According to the time series analysis result set, trend identification is performed. ; in, Indicates that in the target area The trend strength of user behavior within Indicates the target area, is the total number of time points, It is in Time point in target area User stay time, yes The average value of yes The rate of change at a time point is expressed as and The difference, is the normalization constant used The impact of adjustment ensures that comparisons between different regions are comparable, express The average value of , generates the user behavior trend analysis results; According to the user behavior trend analysis results, the weights are adjusted to obtain a dynamically updated user portrait.

2. The smart park intelligent recommendation system based on data analysis according to claim 1 is characterized by: The steps for identifying the activity peak are: Based on the monitoring data in the smart park, collect user movement and residence time data, count the number of users at each time point, and obtain the user density time series; Apply the moving average algorithm to the user density time series to generate a smoothed user density series; Detecting activity peaks based on the smoothed user density sequence, ; in, is the activity peak detection data, It's time point The user density represents the number of users at each time point. is the average user density, reflecting the normal level of the overall user density. It's time point The weights are adjusted according to the importance of the different time periods to optimize the calculation flexibility. is the total number of time points; Significant peak time points are extracted from the activity peak detection data and marked to reveal the intensive periods of user activities, and the activity peak recognition results are output.

3. The smart park intelligent recommendation system based on data analysis according to claim 2 is characterized by: The steps for acquiring the user behavior pattern data are as follows: Call the activity peak recognition result, calculate the activity period estimation result, analyze the interval of the peak time period, obtain the activity period estimation data set, and calculate the new collection frequency. ; in, is the new acquisition frequency, is the average activity period during peak hours, calculated as the average of the intervals during all peak hours, and are the shortest and longest time intervals among all activity cycles, is the adjustment coefficient, which is used to optimize the acquisition frequency and obtain the adjusted acquisition frequency data set; Calling the adjusted collection frequency data set to collect user activity data to obtain an updated user activity data set; Using the updated user activity data set, calculate the user's stay frequency and movement path in different time periods, and conduct preliminary behavior pattern analysis. ; in, is the weighted average behavior index, Indicates Behavioral data of each user, including dwell time and movement frequency, is the behavior weight factor, which is allocated according to the importance of user behavior. is the total number of users, and we get a preliminary set of user behavior pattern analysis results; Through the preliminary user behavior pattern analysis result set, weighted and normalized processing is performed, ; in, It is The number of occurrences of the class behavior pattern, is the weight, determined by the importance or frequency of user behavior, is the smoothing parameter used to adjust the effect of the standard deviation, is the standard deviation of the behavior pattern, It is the total category of behavior patterns, and obtains user behavior pattern data .

4. The smart park intelligent recommendation system based on data analysis according to claim 1 is characterized by: The steps for evaluating the user's response to the current activity are: Extracting the user's behavior log and preference settings from the dynamically updated user portrait to obtain user behavior and preference data; Based on the user behavior and preference data, calculate the user's responsiveness to the activity, ； in, is the user's responsiveness to the target activity, Indicates the user's The intensity of preference for class activities, Indicates The attractiveness of similar activities is analyzed based on the content of the activities and historical user engagement. is the exponential weight of the activity's attractiveness, enhancing the impact of the activity's attractiveness in the responsiveness calculation, Is the user and activity type The distance or mismatch between is the attenuation factor, used to adjust for the effect of distance or mismatch, Indicates the number of activities; The calculated Combined with user portraits, the user's overall reaction to the activity is evaluated to obtain comprehensive response evaluation data.

5. The smart park intelligent recommendation system based on data analysis according to claim 4 is characterized by: The steps for obtaining the activity recommendation list are as follows: Collect user interaction data on activities through the user interface and calculate the user preference weight for each activity: ； in, is the updated value of the user preference, is the user's feedback rating for the nth activity, is the frequency of interaction, indicating the popularity of the activity. is the number of days since the event was launched, used to adjust for the effect of preferences changing over time. is a constant, get the user preference update data, Indicates the number of activities; According to the comprehensive response evaluation data and user preference update data, each activity recommendation score is recalculated and the activity list is sorted, ; in, is the target activity recommendation score, Is the user's Response score for each activity, is the updated activity weight, Indicates the number of activities and generates optimized recommendation list data; Synchronize the optimized recommendation list data to the user interface to ensure that activities with high scores are displayed first in the activity recommendation list.

6. The smart park intelligent recommendation system based on data analysis according to claim 1 is characterized by: The steps for obtaining the intelligent recommendation solution are as follows: Collect user feedback on park facilities, obtain original data on facility satisfaction, and calculate facility satisfaction scores. ; in, is the satisfaction score of the facility, It is User ratings for facilities is the time since the rating, is the time decay coefficient, which is used to adjust the weight of old ratings, is a constant, Generate comprehensive satisfaction analysis data for the total number of facilities; Evaluate the performance of each facility based on the comprehensive satisfaction analysis data, determine whether adjustments are needed based on the user expectation match, and calculate the adjustment recommendation priority. ; in, is the recommended priority of the target facility, is the satisfaction score of the facility, is the average of the satisfaction scores, is the standard deviation of satisfaction scores; according to , dynamically update the activity recommendation list to match user expectations and obtain an intelligent recommendation solution.

Images