A water quality trend prediction and decision support system based on big data analysis

Abstract

The present invention provides a water quality trend prediction and decision support system based on big data analysis, comprising: a monitoring module for real-time monitoring of historical water quality data of aquaculture ponds within a recent preset time period; a prediction module for predicting future water quality trend data of the aquaculture pond based on historical water quality data, based on a pre-mined water quality prediction knowledge set; and a support module for providing visual decision support for users to control the aquaculture ponds based on future water quality trend data. The present invention connects to pre-installed water quality monitoring equipment within the aquaculture area to monitor the historical water quality data of the aquaculture ponds within a recent preset time period in real time. Based on the pre-mined water quality prediction knowledge set, the system then predicts future water quality trend data of the aquaculture ponds based on historical water quality data. Based on future water quality trend data, the system provides visual decision support for users to control the aquaculture ponds, providing efficient and scientific decision support for aquaculture managers, enabling them to make timely water quality adjustments and improve aquaculture results.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and in particular to a water quality trend prediction and decision support system based on big data analysis. Background Art

[0002] With the rapid development of modern aquaculture, water quality management in aquaculture ponds has become a critical factor in protecting the aquaculture environment, improving aquaculture efficiency, and ensuring the quality of aquatic products. Changes in water quality directly impact the growth, reproduction, and health of aquatic organisms. Therefore, timely and accurate monitoring and prediction of water quality changes in aquaculture ponds has become a crucial task in aquaculture management.

[0003] Traditional water quality monitoring usually relies on regular manual testing, but manual regular testing has problems such as insufficient monitoring cycles and poor real-time performance. It cannot fully and accurately reflect the trend of water quality changes in the aquaculture pond. As a result, when aquaculture managers encounter abnormal water quality, it is difficult to make appropriate adjustments in time, which may affect the aquaculture effect.

[0004] Therefore, a solution is urgently needed. Summary of the Invention

[0005] One of the purposes of the present invention is to provide a water quality trend prediction and decision support system based on big data analysis. By connecting to the water quality monitoring equipment deployed in advance in the aquaculture, the historical water quality data of the aquaculture pond within the recent preset time is monitored in real time. There is no need for manual regular water quality monitoring, so that the monitored historical water quality data can comprehensively and accurately reflect the trend of water quality changes in the aquaculture pond. Based on the pre-mined water quality prediction knowledge set, the future water quality trend data of the aquaculture pond is predicted according to the historical water quality data. Based on the future water quality trend data, visual decision support is provided for users to regulate the aquaculture pond, providing efficient and scientific decision support for aquaculture managers, enabling them to make water quality adjustments in a timely manner and improve aquaculture results.

[0006] The present invention provides a water quality trend prediction and decision support system based on big data analysis, including:

[0007] The monitoring module is used to monitor the historical water quality data of the aquaculture pond in real time within the recent preset time;

[0008] The prediction module is used to predict the future water quality trend data of the aquaculture pond based on the pre-mined water quality prediction knowledge set and historical water quality data;

[0009] Support module, used to provide visual decision support for users to regulate aquaculture ponds based on future water quality trend data;

[0010] The steps for pre-mining the water quality prediction knowledge set are as follows:

[0011] Obtaining big data on water quality;

[0012] Analyze and mine water quality prediction knowledge sets from water quality big data.

[0013] Optionally, the support module provides visual decision support for users to regulate aquaculture ponds based on future water quality trend data, including:

[0014] Generate a visualization model based on future water quality trend data and provide it to users for viewing;

[0015] Based on the historical behavior information of users viewing the visualization model, predict the user's decision intention for pond regulation;

[0016] Based on decision intent, the visualization model viewed by users is optimized in a targeted manner.

[0017] Optionally, the support module performs targeted optimization on the visualization model viewed by the user based on the decision intent, including:

[0018] Determine multiple value data search templates corresponding to the decision intent from a value data search template library;

[0019] Based on each value data search template, multiple value data are searched in the visualization model;

[0020] A target local model is separated from the visualization model; wherein the target local model only includes the visualization area of ​​all value data in the visualization model;

[0021] Based on the hierarchical division rules, each valuable data is divided into value data sets of different levels;

[0022] Generates the continued traversal triggering constraints of the value data set;

[0023] Create interactive mechanisms that support user-triggered decision-making;

[0024] Assign the decision interaction mechanism to the target local model to obtain a targeted optimized visualization model;

[0025] The decision interaction mechanism includes:

[0026] When triggered by the user, each value data set is traversed in order from the largest to the smallest level;

[0027] Each time a traversal is completed, the user is interactively guided to view each value data in the traversed value data set. When the interaction between the user and the target local model meets the traversal trigger constraint, the traversal continues to the next value data set.

[0028] Optionally, the hierarchical division rules include:

[0029] Determine multiple decision sub-thoughts corresponding to the decision intention from the decision sub-thought library;

[0030] Associate and pair each value data with each decision sub-idea;

[0031] When the decision sub-thought associated with the value data is unique, the thought level of the decision sub-thought associated with the value data is used as the level of the value data;

[0032] When the decision sub-thoughts associated with the value data are not unique, the maximum thought level of the decision sub-thoughts associated with the value data is used as the level of the value data.

[0033] Optionally, the generation of the trigger constraint for continued traversal of the value data set includes:

[0034] Constraint 1: The user has sequentially viewed more than the target threshold of value data in the value data set, and the maximum matching degree between the viewing rhythm of the value data in the value data set and each standard rhythm in the standard rhythm library exceeds the matching threshold; the target threshold is the upward rounded value of the product of the proportion of the level of the value data set in the proportion value library and the total number of value data in the value data set;

[0035] or,

[0036] Constraint 2: The user has checked in sequence that the data type set of the value data in the value data set matches the standard data type set corresponding to the level of the value data set in the standard type set.

[0037] Optionally, the interactively guiding the user to view each value data in the traversed value data set includes:

[0038] Plan N adjacent time intervals on the future time axis; the starting time node of the future time axis is the moment when the value data set is traversed; the i-th time interval on the future time axis meets the time interval constraint; i = 1, 2, 3, ..., N; N is the value corresponding to the level of the traversed value data set in the time interval quantity library;

[0039] Based on the future timeline completed within the planned time interval, interactively guide users to view each value data in the traversed value data set;

[0040] The time interval constraints include:

[0041] When i is a unique singular number, the time length of the i-th time interval on the future timeline is the value corresponding to i in the first time length library; the i-th time interval is assigned a first interaction working mechanism, including: taking the value data that the user has not viewed in the traversed value data set as the target to be viewed, continuously detecting the target to be viewed that the user actively views and the target to be viewed that the user actively views during the viewing gap of the target to be viewed passively, and if the former is not empty, only the former is determined as the interaction basis; if only the former is empty, the latter is determined as the interaction basis;

[0042] When i is a unique even number, the time length of the i-th time interval on the future timeline is the value corresponding to i in the second time length library; assigning the i-th time interval a second interaction working mechanism, including: taking the value data in the traversed value data set that has at least one association relationship with the interaction basis determined in the previous i-1 time interval and the target to be viewed passively viewed by the user as the target to be guided, and guiding the user to view the target to be guided;

[0043] The maximum length value in the first time length library does not exceed the minimum length value in the second time length library.

[0044] An embodiment of the present invention provides a water quality trend prediction and decision support method based on big data analysis, including:

[0045] Real-time monitoring of historical water quality data of the aquaculture pond within the recent preset time;

[0046] Based on the pre-mined water quality prediction knowledge set and historical water quality data, the future water quality trend data of the aquaculture pond is predicted;

[0047] Based on future water quality trend data, users can be visually supported in making decisions on how to regulate aquaculture ponds;

[0048] The steps for pre-mining the water quality prediction knowledge set are as follows:

[0049] Obtaining big data on water quality;

[0050] Analyze and mine water quality prediction knowledge sets from water quality big data.

[0051] Optionally, the visual decision support for users to regulate aquaculture ponds based on future water quality trend data includes:

[0052] Generate a visualization model based on future water quality trend data and provide it to users for viewing;

[0053] Based on the historical behavior information of users viewing the visualization model, predict the user's decision intention for pond regulation;

[0054] Based on decision intent, the visualization model viewed by users is optimized in a targeted manner.

[0055] Optionally, the targeted optimization of the visualization model viewed by the user based on the decision intent includes:

[0056] Determine multiple value data search templates corresponding to the decision intent from a value data search template library;

[0057] Based on each value data search template, multiple value data are searched in the visualization model;

[0058] A target local model is separated from the visualization model; wherein the target local model only includes the visualization area of ​​all value data in the visualization model;

[0059] Based on the hierarchical division rules, each valuable data is divided into value data sets of different levels;

[0060] Generates the continued traversal triggering constraints of the value data set;

[0061] Create interactive mechanisms that support user-triggered decision-making;

[0062] Assign the decision interaction mechanism to the target local model to obtain a targeted optimized visualization model;

[0063] The decision interaction mechanism includes:

[0064] When triggered by the user, each value data set is traversed in order from the largest to the smallest level;

[0065] Each time a traversal is completed, the user is interactively guided to view each value data in the traversed value data set. When the interaction between the user and the target local model meets the traversal trigger constraint, the traversal continues to the next value data set.

[0066] Optionally, the hierarchical division rules include:

[0067] Determine multiple decision sub-thoughts corresponding to the decision intention from the decision sub-thought library;

[0068] Associate and pair each value data with each decision sub-idea;

[0069] When the decision sub-thought associated with the value data is unique, the thought level of the decision sub-thought associated with the value data is used as the level of the value data;

[0070] When the decision sub-thoughts associated with the value data are not unique, the maximum thought level of the decision sub-thoughts associated with the value data is used as the level of the value data.

[0071] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0072] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0074] Figure 1 Schematic diagram of a water quality trend prediction and decision support system based on big data analysis in an embodiment of the present invention;

[0075] Figure 2 Schematic diagram of a water quality trend prediction and decision support method based on big data analysis in an embodiment of the present invention. DETAILED DESCRIPTION

[0076] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0077] The embodiment of the present invention provides a water quality trend prediction and decision support system based on big data analysis, such as Figure 1 As shown, including:

[0078] Monitoring module 1 is used to monitor the historical water quality data of the aquaculture pond in real time within the recent preset time;

[0079] Prediction module 2 is used to predict the future water quality trend data of the aquaculture pond based on the pre-mined water quality prediction knowledge set and historical water quality data;

[0080] Support module 3 is used to provide visual decision support for users to regulate aquaculture ponds based on future water quality trend data;

[0081] The steps for pre-mining the water quality prediction knowledge set are as follows:

[0082] Obtaining big data on water quality;

[0083] Analyze and mine water quality prediction knowledge sets from water quality big data.

[0084] In the above technical solution, historical water quality data is collected by connecting water quality monitoring equipment (such as water temperature, pH value, dissolved oxygen, ammonia nitrogen and other water quality parameter monitoring sensors) deployed in advance in the aquaculture; the most recent preset time can be, for example, the past 24 hours, the past week, etc.; the water quality prediction knowledge set contains a large number of water quality change patterns; therefore, based on the water quality prediction knowledge set, the future water quality trend data of the aquaculture pond can be predicted according to the historical water quality data; based on the prediction results, visual decision support suggestions are provided to users to help aquaculture managers take timely measures to regulate the water quality of the aquaculture pond and reduce the impact of abnormal water quality on the aquaculture environment. For example, it shows the water quality of each pond in the future. The system can generate a curve of water quality parameter changes and mark possible water quality risks (such as too low pH value, too little dissolved oxygen, etc.); when pre-mining the water quality prediction knowledge set, a large amount of water quality data is collected from multiple monitoring points in the aquaculture pond, historical water quality records and other relevant environmental data. These data may come from sensors, manual sampling or external data sources. By integrating these data, water quality big data is obtained. Through data mining techniques such as cluster analysis, association rule analysis, time series analysis, etc., the system can identify the relationship and rules between water quality parameters, that is, to mine the water quality change rules, such as: the time series relationship of water quality parameters, the correlation between different water quality parameters, the periodic characteristics of water quality changes, etc.

[0085] This application connects to the water quality monitoring equipment pre-deployed in the aquaculture to monitor the historical water quality data of the aquaculture pond in the recent preset time in real time, without the need for manual water quality monitoring on a regular basis, so that the monitored historical water quality data can comprehensively and accurately reflect the trend of water quality changes in the aquaculture pond. Based on the pre-mined water quality prediction knowledge set, the future water quality trend data of the aquaculture pond is predicted according to the historical water quality data. Based on the future water quality trend data, visual decision support is provided for users to regulate the aquaculture pond, providing efficient and scientific decision support for aquaculture managers, enabling them to make timely water quality adjustments and improve aquaculture results.

[0086] In one embodiment, the support module 3 provides visual decision support for users to regulate the aquaculture pond based on future water quality trend data, including:

[0087] Generate a visualization model based on future water quality trend data and provide it to users for viewing;

[0088] Based on the historical behavior information of users viewing the visualization model, predict the user's decision intention for pond regulation;

[0089] Based on decision intent, the visualization model viewed by users is optimized in a targeted manner.

[0090] The visualization model is a digital model that visualizes future water quality trend data. Through the visualization model, users can intuitively view future water quality trend data. The visualization model is provided to users for viewing. When users view the visualization model, historical behavior information is generated, such as: the content viewed on the visualization model, the viewing time, the operations performed during the viewing, etc. The historical behavior information reflects how the user wants to regulate the aquaculture pond. Therefore, the decision intention can be predicted based on the historical behavior information. The decision intention is the user's intention on how to regulate the aquaculture pond. After the user's decision intention is predicted, the visualization model viewed by the user is targeted and optimized so that the display of the visualization model is consistent with the user's decision intention, which facilitates the user to quickly make aquaculture pond regulation decisions.

[0091] In one embodiment, the support module 3 performs targeted optimization on the visualization model viewed by the user based on the decision intention, including:

[0092] S1. Determine multiple value data search templates corresponding to decision intentions from a value data search template library;

[0093] In S1, the value data search template library contains value data search templates corresponding to different decision intentions. The value data search templates are templates used by the system to search for value data corresponding to the decision intention in the visualization model. For example, if the decision intention is to adjust the water temperature of a sub-tank in the breeding pond, the current water temperature data, future water temperature change data, and suitable water temperature data for breeding animals of the sub-tank are searched as multiple value data. The value data is data that has auxiliary value for the user to make breeding pond regulation decisions based on the decision intention.

[0094] S2. Based on each value data search template, multiple value data are searched in the visualization model;

[0095] In S2, based on each value data search template, the value data can be searched in the visualization model;

[0096] S3. Segmenting a target local model from the visualization model; wherein the target local model only includes the visualization area of ​​all value data in the visualization model;

[0097] In S3, the value data is displayed in a visualization model, which has a visualization area for display. The visualization area of ​​all the value data in the visualization model is combined to form a target local model, which is separated from the visualization model. When the user views the target local model, they can only view all the value data, which facilitates their immersion in the aquaculture pond regulation decision-making under the decision-making intention;

[0098] S4. Based on the hierarchical division rules, each value data is divided into value data sets of different levels;

[0099] S5. Generate the trigger constraint for the continued traversal of the value data set;

[0100] S6. Create a decision-making interaction mechanism that supports user triggers;

[0101] S7. Assign the decision interaction mechanism to the target local model to obtain a targeted optimized visualization model;

[0102] In S7, after the decision interaction mechanism is assigned to the target local model, the user can choose to trigger the decision interaction mechanism when viewing the target local model. Once triggered, the system will execute the decision interaction mechanism; the target local model assigned with the decision interaction mechanism will be used as a targeted optimized visualization model;

[0103] The decision interaction mechanism includes:

[0104] When triggered by the user, each value data set is traversed in order from the largest to the smallest level;

[0105] Each time a traversal is completed, the user is interactively guided to view each value data in the traversed value data set. When the interaction between the user and the target local model meets the traversal trigger constraint, the traversal continues to the next value data set.

[0106] In the decision-making interaction mechanism, each value data set is traversed in sequence according to the size of the hierarchy. Each time a traversal is completed, the user is interactively guided to view the various value data in the traversed value data set, so that the user can view the value data in the value data set with a larger hierarchy first, thereby improving the efficiency of the aquaculture pond regulation decision under its decision-making intention; when the interaction between the user and the target local model meets the traversal trigger constraint, the next value data set will be traversed, and the timing of the traversal switch will be accurately determined, which reduces the system's decision-making interaction resources and improves the system's decision-making interaction efficiency.

[0107] In one embodiment, the hierarchical division rules include:

[0108] Determine multiple decision sub-thoughts corresponding to the decision intention from the decision sub-thought library;

[0109] Associate and pair each value data with each decision sub-idea;

[0110] When the decision sub-thought associated with the value data is unique, the thought level of the decision sub-thought associated with the value data is used as the level of the value data;

[0111] When the decision sub-thoughts associated with the value data are not unique, the maximum thought level of the decision sub-thoughts associated with the value data is used as the level of the value data.

[0112] There are decision sub-thoughts corresponding to different decision intentions in the decision sub-thought library. Decision sub-thoughts are multiple standard thoughts for making decisions on aquaculture pond regulation under decision intentions. For example, if the decision intention is to adjust the water temperature of a sub-pond in the aquaculture pond, first analyze the future water temperature changes of the sub-pond, then analyze the suitable water temperature for the aquaculture animals in the sub-pond, and then analyze whether the future water temperature changes of the sub-pond are suitable for the aquaculture animals in the sub-pond. These analysis steps form decision sub-thoughts. Decision sub-thoughts have thought levels and are divided into D-level, D-1-level, D-2-level, etc. according to the order of their execution, where D is the total number of decision sub-thoughts. Each value data is respectively associated with each decision sub-thought. When the ideas are associated and paired, the value data needed for the execution of the decision sub-idea are associated and paired with the corresponding decision sub-idea. For example, if the decision sub-idea is to analyze the future water temperature changes of the sub-pool, the future water temperature change trend data of the sub-pool is associated and paired with the decision sub-idea. When the decision sub-idea associated and paired with the value data is unique, the idea level of the decision sub-idea associated with the value data is used as the level of the value data. When the decision sub-idea associated and paired with the value data is not unique, in order to ensure that the user can view the necessary value data in a timely manner, the maximum idea level of the decision sub-idea associated with the value data is used as the level of the value data.

[0113] In one embodiment, generating a continued traversal trigger constraint of a value dataset includes:

[0114] Constraint 1: The user has sequentially viewed more than the target threshold of value data in the value data set, and the maximum matching degree between the viewing rhythm of the value data in the value data set and each standard rhythm in the standard rhythm library exceeds the matching threshold; the target threshold is the upward rounded value of the product of the proportion of the level of the value data set in the proportion value library and the total number of value data in the value data set;

[0115] or,

[0116] Constraint 2: The user has checked in sequence that the data type set of the value data in the value data set matches the standard data type set corresponding to the level of the value data set in the standard type set.

[0117] In constraint one, there are proportion values ​​corresponding to different levels in the proportion value library. The larger the level, the more the user is in the initial stage of interacting with the system, the more the system guidance is needed, and the corresponding proportion value is larger; the standard rhythm represents the viewing rhythm of the user carefully following the system guidance to view the value data, for example: the viewing time of each value data is more than 100 seconds, etc.; the matching threshold can be 80%; if the user has viewed more than the target threshold value of value data in the value data set in sequence and the maximum matching degree between the viewing rhythm of the value data in the value data set in sequence and the standard rhythm in the standard rhythm library exceeds the matching threshold, it means that the user is immersed in the guidance of the system, and the idea of ​​​​making decisions on regulating the breeding pond formed under the guidance has basically taken shape, then the guidance can be stopped. In constraint two, the standard type set has standard data type sets corresponding to different levels. The standard data type set is the data type of the value data that the user needs to view at this level. For example, if the level is the largest level 3, the data type of the value data that the user needs to view is the future water temperature change trend of different temperature monitoring points in the sub-pools that need to be temperature-controlled in the aquaculture pond. When the user has viewed the data type set of the value data in the value data set in sequence and it matches the standard data type set corresponding to the level of the value data set in the standard type set, it also means that the user can continue to traverse the next value data set.

[0118] In one embodiment, the interactively guiding the user to view each value data in the traversed value data set includes:

[0119] S51. Plan N adjacent time intervals on the future time axis; the starting time node of the future time axis is the moment when the value dataset is traversed; the i-th time interval on the future time axis meets the time interval constraint; i = 1, 2, 3, ..., N; N is the value corresponding to the level of the traversed value dataset in the time interval quantity library;

[0120] In S51, the time interval quantity library has quantity values ​​corresponding to different levels. The larger the level, the more the user is in the initial stage of interaction with the system, the more refined guidance the system needs, the more time intervals need to be set, and the larger the corresponding quantity value; the time intervals are adjacent to each other on the future time axis and arranged closely. The starting time node of the future time axis is set to the moment when the value data set is traversed. The moment refers to the traversal related to the traversed value data set.

[0121] S52. Based on the future timeline completed during the planned time interval, interactively guide the user to view each value data in the traversed value data set;

[0122] In S52, based on the future timeline completed in the planned time interval, when interactively guiding the user to view each value data in the traversed value data set, a time change pointer is set on the timeline, and moves from the starting time node as time passes. Each time it moves into a time interval, the first interactive working mechanism or the second interactive working mechanism assigned to the time interval is executed;

[0123] The time interval constraints include:

[0124] Constraint 3: When i is a unique singular number, the time length of the i-th time interval on the future timeline is the value corresponding to i in the first time length library; assign the i-th time interval a first interaction working mechanism, including: treating the value data that the user has not viewed in the traversed value data set as the target to be viewed, continuously detecting the target to be viewed that the user actively views and the target to be viewed that the user actively views during the gaps between passively viewing the target to be viewed, and if the former is not empty, only the former is determined as the interaction basis; if only the former is empty, the latter is determined as the interaction basis;

[0125] Constraint 4: When i is a unique even number, the length of the i-th time interval on the future timeline is the value corresponding to i in the second time length library; assigning the i-th time interval a second interaction working mechanism, including: treating the value data in the traversed value data set that has at least one association with the interaction basis determined in the previous i-1 time interval and the target to be viewed passively viewed by the user as the target to be guided, and guiding the user to view the target to be guided;

[0126] The maximum length value in the first time length library does not exceed the minimum length value in the second time length library.

[0127] In constraint three, the first time length library contains time length values ​​corresponding to different values ​​of i. The smaller i is, the more time is required to perform the interaction, and the corresponding time length value is larger. Active viewing by the user refers to the user actively viewing the target in the target local model; passive viewing refers to viewing the target under the interactive guidance of the system; the viewing gap of the passive viewing of the target refers to the time gap between the user viewing different targets under the interactive guidance of the system. Taking i as a unique 1 as an example, first determine the corresponding value from the first time length library, such as 300 seconds, then set the time length of the first time interval to 300 seconds. If the user actively views a target, this value is used as the interaction basis for determining the first time interval. If the user does not actively view a target, but actively views the target during the viewing gap of the passive viewing of the target, the target actively viewed by the user during the viewing gap of the passive viewing of the target is used as the interaction basis for determining the first time interval (the target actively viewed by the user can better reflect the user's specific decision-making ideas than the target viewed during the viewing gap, so the former is preferred as the interaction basis). In constraint four, there are time length values ​​corresponding to different i in the first time length library. The smaller i is, the more time is required to perform the interaction, and the corresponding time length value is larger; since the selection of interaction basis takes a shorter time, the maximum length value in the first time length library does not exceed the minimum length value in the second time length library; the association relationship refers to the relationship between the user's subsequent viewing of the value data, which is beneficial to its auxiliary decision-making, and can be set in advance by technical personnel according to actual needs; taking i as the only 2 as an example, first determine the corresponding value such as 250 seconds from the first time length library, then set the time length of the second time interval to 250 seconds, and use the value data in the traversed value data set that has at least one association relationship with the interaction basis determined in the first time interval and the target to be viewed passively viewed by the user as the target to be guided, and guide the user to view the target to be guided.

[0128] When the present application interactively guides the user to view each value data in the traversed value data set, based on the future timeline completed by the planned time interval, the user is interactively guided to view each value data in the traversed value data set, which greatly improves the accuracy and efficiency of the interactive guidance, improves the applicability of the system, and is more humane at the same time; when planning N adjacent time intervals on the future timeline, the i-th time interval on the future timeline meets the time interval constraint, and the time interval constraint can accurately plan what kind of first interactive working mechanism and second interactive working mechanism are most suitable at different times, which greatly improves the accuracy and comprehensiveness of planning time intervals on the future timeline.

[0129] The embodiment of the present invention provides a water quality trend prediction and decision support method based on big data analysis, such as Figure 2Shown, including:

[0130] Step 1: Real-time monitoring of historical water quality data of the aquaculture pond within the recent preset time;

[0131] Step 2: Based on the pre-mined water quality prediction knowledge set and historical water quality data, predict the future water quality trend data of the aquaculture pond;

[0132] Step 3: Based on future water quality trend data, provide visual decision support for users to regulate aquaculture ponds;

[0133] The steps for pre-mining the water quality prediction knowledge set are as follows:

[0134] Obtaining big data on water quality;

[0135] Analyze and mine water quality prediction knowledge sets from water quality big data.

[0136] The visual decision support for users to regulate aquaculture ponds based on future water quality trend data includes:

[0137] Generate a visualization model based on future water quality trend data and provide it to users for viewing;

[0138] Based on the historical behavior information of users viewing the visualization model, predict the user's decision intention for pond regulation;

[0139] Based on decision intent, the visualization model viewed by users is optimized in a targeted manner.

[0140] The targeted optimization of the visualization model viewed by the user based on the decision intent includes:

[0141] Determine multiple value data search templates corresponding to the decision intent from a value data search template library;

[0142] Based on each value data search template, multiple value data are searched in the visualization model;

[0143] A target local model is separated from the visualization model; wherein the target local model only includes the visualization area of ​​all value data in the visualization model;

[0144] Based on the hierarchical division rules, each valuable data is divided into value data sets of different levels;

[0145] Generates the continued traversal triggering constraints of the value data set;

[0146] Create interactive mechanisms that support user-triggered decision-making;

[0147] Assign the decision interaction mechanism to the target local model to obtain a targeted optimized visualization model;

[0148] The decision interaction mechanism includes:

[0149] When triggered by the user, each value data set is traversed in order from the largest to the smallest level;

[0150] Each time a traversal is completed, the user is interactively guided to view each value data in the traversed value data set. When the interaction between the user and the target local model meets the traversal trigger constraint, the traversal continues to the next value data set.

[0151] The hierarchical division rules include:

[0152] Determine multiple decision sub-thoughts corresponding to the decision intention from the decision sub-thought library;

[0153] Associate and pair each value data with each decision sub-idea;

[0154] When the decision sub-thought associated with the value data is unique, the thought level of the decision sub-thought associated with the value data is used as the level of the value data;

[0155] When the decision sub-thoughts associated with the value data are not unique, the maximum thought level of the decision sub-thoughts associated with the value data is used as the level of the value data.

[0156] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A water quality trend prediction and decision support system based on big data analysis, characterized in that: include: The monitoring module is used to monitor the historical water quality data of the aquaculture pond in real time within the recent preset time; The prediction module is used to predict the future water quality trend data of the aquaculture pond based on the pre-mined water quality prediction knowledge set and historical water quality data; Support module, used to provide visual decision support for users to regulate aquaculture ponds based on future water quality trend data; The steps for pre-mining the water quality prediction knowledge set are as follows: Obtaining big data on water quality; Analyze and mine water quality prediction knowledge sets from water quality big data; The support module provides visual decision support for users to regulate aquaculture ponds based on future water quality trend data, including: Generate a visualization model based on future water quality trend data and provide it to users for viewing; Based on the historical behavior information of users viewing the visualization model, predict the user's decision intention for pond regulation; Based on decision-making intentions, targeted optimization is performed on the visualization model viewed by users; The support module performs targeted optimization on the visualization model viewed by the user based on the decision intention, including: Determine multiple value data search templates corresponding to the decision intent from a value data search template library; Based on each value data search template, multiple value data are searched in the visualization model; A target local model is separated from the visualization model; wherein the target local model only includes the visualization area of ​​all value data in the visualization model; Based on the hierarchical division rules, each valuable data is divided into value data sets of different levels; Generates the continued traversal triggering constraints of the value data set; Create interactive mechanisms that support user-triggered decision-making; Assign the decision interaction mechanism to the target local model to obtain a targeted optimized visualization model; The decision interaction mechanism includes: When triggered by the user, each value data set is traversed in order from the largest to the smallest level; Each time a traversal is completed, the user is interactively guided to view each value data in the traversed value data set. When the interaction between the user and the target local model meets the traversal trigger constraint, the next value data set is traversed. The interactive guidance of the user to view each value data in the traversed value data set includes: Plan N adjacent time intervals on the future time axis; the starting time node of the future time axis is the moment when the value data set is traversed; the i-th time interval on the future time axis meets the time interval constraint; i = 1, 2, 3, ..., N; N is the value corresponding to the level of the traversed value data set in the time interval quantity library; Based on the future timeline completed within the planned time interval, interactively guide users to view each value data in the traversed value data set; The time interval constraints include: When i is a unique singular number, the time length of the i-th time interval on the future timeline is the value corresponding to i in the first time length library; the i-th time interval is assigned a first interaction working mechanism, including: taking the value data that the user has not viewed in the traversed value data set as the target to be viewed, continuously detecting the target to be viewed that the user actively views and the target to be viewed that the user actively views during the viewing gap of the target to be viewed passively, and if the former is not empty, only the former is determined as the interaction basis; if only the former is empty, the latter is determined as the interaction basis; When i is a unique even number, the time length of the i-th time interval on the future timeline is the value corresponding to i in the second time length library; assigning the i-th time interval a second interaction working mechanism, including: taking the value data in the traversed value data set that has at least one association relationship with the interaction basis determined in the previous i-1 time interval and the target to be viewed passively viewed by the user as the target to be guided, and guiding the user to view the target to be guided; The maximum length value in the first time length library does not exceed the minimum length value in the second time length library.

2. The water quality trend prediction and decision support system based on big data analysis according to claim 1, characterized in that: The hierarchical division rules include: Determine multiple decision sub-thoughts corresponding to the decision intention from the decision sub-thought library; Associate and pair each value data with each decision sub-idea; When the decision sub-thought associated with the value data is unique, the thought level of the decision sub-thought associated with the value data is used as the level of the value data; When the decision sub-thoughts associated with the value data are not unique, the maximum thought level of the decision sub-thoughts associated with the value data is used as the level of the value data.

3. The water quality trend prediction and decision support system based on big data analysis according to claim 1, characterized in that: The continued traversal triggering constraints for generating the value data set include: Constraint 1: The user has sequentially viewed more than the target threshold of value data in the value data set, and the maximum matching degree between the viewing rhythm of the value data in the value data set and each standard rhythm in the standard rhythm library exceeds the matching threshold; the target threshold is the upward rounded value of the product of the proportion of the level of the value data set in the proportion value library and the total number of value data in the value data set; or, Constraint 2: The user has checked in sequence that the data type set of the value data in the value data set matches the standard data type set corresponding to the level of the value data set in the standard type set.

4. A water quality trend prediction and decision support method based on big data analysis, characterized in that: include: Real-time monitoring of historical water quality data of the aquaculture pond within the recent preset time; Based on the pre-mined water quality prediction knowledge set and historical water quality data, the future water quality trend data of the aquaculture pond is predicted; Based on future water quality trend data, users can be visually supported in making decisions on how to regulate aquaculture ponds; The steps for pre-mining the water quality prediction knowledge set are as follows: Obtaining big data on water quality; Analyze and mine water quality prediction knowledge sets from water quality big data; The visual decision support for users to regulate aquaculture ponds based on future water quality trend data includes: Generate a visualization model based on future water quality trend data and provide it to users for viewing; Based on the historical behavior information of users viewing the visualization model, predict the user's decision intention for pond regulation; Based on decision-making intentions, targeted optimization is performed on the visualization model viewed by users; The targeted optimization of the visualization model viewed by the user based on the decision intention includes: Determine multiple value data search templates corresponding to the decision intent from a value data search template library; Based on each value data search template, multiple value data are searched in the visualization model; A target local model is separated from the visualization model; wherein the target local model only includes the visualization area of ​​all value data in the visualization model; Based on the hierarchical division rules, each valuable data is divided into value data sets of different levels; Generates the continued traversal triggering constraints of the value data set; Create interactive mechanisms that support user-triggered decision-making; Assign the decision interaction mechanism to the target local model to obtain a targeted optimized visualization model; The decision interaction mechanism includes: When triggered by the user, each value data set is traversed in order from the largest to the smallest level; Each time a traversal is completed, the user is interactively guided to view each value data in the traversed value data set. When the interaction between the user and the target local model meets the traversal trigger constraint, the next value data set is traversed. The interactive guidance of the user to view each value data in the traversed value data set includes: Plan N adjacent time intervals on the future time axis; the starting time node of the future time axis is the moment when the value data set is traversed; the i-th time interval on the future time axis meets the time interval constraint; i = 1, 2, 3, ..., N; N is the value corresponding to the level of the traversed value data set in the time interval quantity library; Based on the future timeline completed within the planned time interval, interactively guide users to view each value data in the traversed value data set; The time interval constraints include: When i is a unique singular number, the time length of the i-th time interval on the future timeline is the value corresponding to i in the first time length library; the i-th time interval is assigned a first interaction working mechanism, including: taking the value data that the user has not viewed in the traversed value data set as the target to be viewed, continuously detecting the target to be viewed that the user actively views and the target to be viewed that the user actively views during the viewing gap of the target to be viewed passively, and if the former is not empty, only the former is determined as the interaction basis; if only the former is empty, the latter is determined as the interaction basis; When i is a unique even number, the time length of the i-th time interval on the future timeline is the value corresponding to i in the second time length library; assigning the i-th time interval a second interaction working mechanism, including: taking the value data in the traversed value data set that has at least one association relationship with the interaction basis determined in the previous i-1 time interval and the target to be viewed passively viewed by the user as the target to be guided, and guiding the user to view the target to be guided; The maximum length value in the first time length library does not exceed the minimum length value in the second time length library.

5. The water quality trend prediction and decision support method based on big data analysis according to claim 4 is characterized in that: The hierarchical division rules include: Determine multiple decision sub-thoughts corresponding to the decision intention from the decision sub-thought library; Associate and pair each value data with each decision sub-idea; When the decision sub-thought associated with the value data is unique, the thought level of the decision sub-thought associated with the value data is used as the level of the value data; When the decision sub-thoughts associated with the value data are not unique, the maximum thought level of the decision sub-thoughts associated with the value data is used as the level of the value data.

Images