CN114781952A - Risk early warning method for epidemic disease prevention and control in animal husbandry - Google Patents

Risk early warning method for epidemic disease prevention and control in animal husbandry Download PDF

Info

Publication number
CN114781952A
CN114781952A CN202210714501.0A CN202210714501A CN114781952A CN 114781952 A CN114781952 A CN 114781952A CN 202210714501 A CN202210714501 A CN 202210714501A CN 114781952 A CN114781952 A CN 114781952A
Authority
CN
China
Prior art keywords
cow
cows
vector
normal
social network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202210714501.0A
Other languages
Chinese (zh)
Other versions
CN114781952B (en
Inventor
刘志阳
张厚林
张勇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jining Rencheng District Animal Husbandry And Veterinary Development Center Jining Rencheng District Animal Disease Prevention And Control Center Jining Rencheng District Animal Health And Quarantine Center
Original Assignee
Jining Rencheng District Animal Husbandry And Veterinary Development Center Jining Rencheng District Animal Disease Prevention And Control Center Jining Rencheng District Animal Health And Quarantine Center
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Jining Rencheng District Animal Husbandry And Veterinary Development Center Jining Rencheng District Animal Disease Prevention And Control Center Jining Rencheng District Animal Health And Quarantine Center filed Critical Jining Rencheng District Animal Husbandry And Veterinary Development Center Jining Rencheng District Animal Disease Prevention And Control Center Jining Rencheng District Animal Health And Quarantine Center
Priority to CN202210714501.0A priority Critical patent/CN114781952B/en
Publication of CN114781952A publication Critical patent/CN114781952A/en
Application granted granted Critical
Publication of CN114781952B publication Critical patent/CN114781952B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/01—Social networking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/02—Agriculture; Fishing; Mining
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
    • Y02A40/70—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in livestock or poultry

Abstract

The invention relates to the technical field of data processing, in particular to a risk early warning method for epidemic disease prevention and control in animal husbandry. The method comprises the following steps: acquiring attribute vectors and location vectors of each normal cow and each diseased cow; connecting all the contact cows to construct a cow social network; dividing all normal cows in the social network of cows into dangerous cows and safe cows according to the connection relation in the social network of cows; updating the attribute vector and the location vector of each dangerous cow and each safe cow in the cow social network; the target attribute vector and the target location vector of the sick dairy cow are obtained, the epidemic disease risk level of the normal dairy cow is obtained by combining the updated attribute vector and location vector of each normal dairy cow, early warning is timely carried out based on the epidemic disease risk level, reliability in the data analysis process is improved, and the epidemic disease prevention and control efficiency is improved.

Description

Risk early warning method for epidemic disease prevention and control in animal husbandry
Technical Field
The invention relates to the technical field of data processing, in particular to a risk early warning method for epidemic disease prevention and control in animal husbandry.
Background
China is a big animal husbandry country with rich natural resources, most of cattle, sheep and other livestock love ecological breeding, and the breeding mode includes two categories of captive breeding and stocking; livestock bred in a natural grassland not only eat more nutrition and health, but also can exercise the body of the livestock, so that the meat quality is more compact, and the livestock breeding method is a more humane breeding mode; however, the uncontrollable factors in the natural grassland environment are too many, and the livestock can be sick due to factors such as soil pollution, germs carried by other organisms, environment moisture and the like, and even can die in a large area when the livestock is serious, so that great economic loss is caused. Therefore, health tests are required for each individual animal, but the difficulty of health tests for stocked animals is greater than the difficulty of housing.
In most of the prior art, natural environment detection and livestock physical examination are carried out manually and regularly, and once an individual livestock is found to have infectious diseases, physical examination judgment needs to be carried out on the whole group; however, the method of using the artificial periodic physical examination requires a great amount of manpower and material resources, and may not find the individual suffering from the disease in time, thereby increasing the risk of outbreak of the whole group suffering from the disease, and the detection error is large and the timeliness and efficiency are low.
Disclosure of Invention
In order to solve the technical problem, the invention aims to provide a risk early warning method for epidemic prevention and control in animal husbandry, which comprises the following steps:
acquiring an attribute vector and a location vector of each cow, wherein the cows comprise normal cows and sick cows; acquiring corresponding regional marginality and regional fixity based on the location vector of each cow;
connecting all dairy cows with contact to construct a social network of the dairy cows, wherein the contact is that the distance between any two dairy cows is smaller than a preset range; dividing all normal cows in the cow social network into dangerous cows and safe cows according to the connection relation in the cow social network;
updating the location vector according to the regional marginality and the regional stationarity corresponding to each normal cow in the social network of the cows to obtain a social environment vector; updating the attribute vector of each dangerous cow in the cow social network to obtain a social attribute vector, and updating the attribute vector of each safe cow in the cow social network to obtain a social perception vector;
acquiring a target attribute vector and a target location vector of the sick dairy cows, and calculating attribute similarity between each normal dairy cow and the sick dairy cows as a first similarity, wherein the attribute similarity comprises cosine similarity between a social attribute vector and the target attribute vector and cosine similarity between a social perception vector and the target attribute vector; calculating cosine similarity between the social environment vector of each normal cow and the target location vector to be second similarity, obtaining the epidemic risk level of each normal cow based on the first similarity and the second similarity of each normal cow, and timely early warning based on the epidemic risk level.
Preferably, the step of dividing all normal cows in the cow social network into dangerous cows and safe cows according to the connection relationship in the cow social network includes:
acquiring a contact risk route corresponding to each sick cow based on the cow social network, wherein the contact risk route is composed of all normal cows connected with the sick cows, and the connection comprises direct connection and indirect connection; normal cows on the contact risk route are dangerous cows, and normal cows not on the contact risk route are safe cows.
Preferably, the step of connecting all the contact cows to construct a social network of cows includes:
taking each cow as a node, and connecting the nodes corresponding to two cows when any two cows are contacted; and connecting all the nodes corresponding to the contacted cows to obtain a social network of the cows.
Preferably, the step of updating the location vector according to regional marginality and regional stationarity corresponding to each normal cow in the social network of cows to obtain a social environment vector includes:
selecting nodes corresponding to any normal cow in the cow social network as points to be updated, obtaining a neighborhood node set according to all nodes directly connected with the points to be updated, calculating to obtain corresponding weights according to regional edge lines and regional fixity corresponding to each node in the neighborhood node set, and obtaining a social environment vector of the points to be updated based on the weighted summation of the weights of all nodes in the neighborhood node set and corresponding location vectors.
Preferably, the step of updating the attribute vector of each dangerous cow in the cow social network to obtain a social attribute vector includes:
marking the sick cows in the cow social network as 0 layer, wherein the dangerous cows directly connected with the sick cows are 1 layer, the dangerous cows directly connected with the dangerous cows on the 1 layer are 2 layers, and so on, and marking the number of layers of all the dangerous cows in the cow social network;
selecting a node corresponding to any dangerous cow in the cow social network as a point to be processed, acquiring all adjacent nodes directly connected with the point to be processed, wherein the number of layers in all adjacent nodes is smaller than that of the point to be processed, the adjacent nodes are lower-layer neighborhood points, acquiring a weight corresponding to each lower-layer neighborhood point, and acquiring a social attribute vector of the point to be processed based on weighted summation between the weights of all the lower-layer neighborhood points and corresponding attribute vectors.
Preferably, the step of updating the attribute vector of each safe cow in the cow social network to obtain a social perception vector includes:
selecting a node corresponding to any safe cow in the cow social network as a target point, obtaining a neighborhood node set according to all nodes directly connected with the target point, obtaining the weight of each node in the neighborhood node set, and obtaining the social perception vector of the target point based on the weighted summation of the weights of all nodes in the neighborhood node set and corresponding attribute vectors.
Preferably, the step of obtaining the epidemic risk rating of the normal cows based on the first similarity and the second similarity of each normal cow comprises:
the epidemic risk levels are:
Figure 339749DEST_PATH_IMAGE002
wherein, the first and the second end of the pipe are connected with each other,
Figure DEST_PATH_IMAGE003
representing the epidemic risk level corresponding to the normal cow;
Figure 368141DEST_PATH_IMAGE004
representing a first similarity of the normal cow;
Figure DEST_PATH_IMAGE005
representing a second similarity of the normal cow;
Figure 23244DEST_PATH_IMAGE006
represents an adjustment coefficient;
Figure DEST_PATH_IMAGE007
indicating rounding up.
Preferably, the step of obtaining the attribute vector of each cow includes:
labeling the cow as a diseased cow or a normal cow as a label of the cow, elements in an attribute vector of the cow including: label, age, breed, and number of diseases.
Preferably, the obtaining of the location vector of each cow includes:
dividing a grassland area into a plurality of sub-areas, acquiring activity place information of each cow in a historical time period, constructing a corresponding place thermodynamic diagram according to the activity place information of each cow, and acquiring the heat of each sub-area according to the place thermodynamic diagrams;
and selecting a plurality of sub-areas with the highest heat degree in all the sub-areas to sequence to obtain corresponding heat degree sequences, wherein the heat degree sequences are the place vectors corresponding to the cows.
Preferably, the step of obtaining corresponding regional marginality and regional stationarity based on the location vector of each cow includes:
acquiring the outermost sub-region in the grassland region as an edge region, and counting the number of the edge regions in the location vector corresponding to each cow, wherein the ratio of the number of the edge regions to the number of all the sub-regions is the regional marginality of the cow;
and acquiring the number of 0 element values in the location vector of the cow, and obtaining the regional fixity of the cow based on the number of 0 element values and the number of all elements in the location vector.
The invention has the following beneficial effects: the vector of each normal cow is updated based on the connection relationship between the cow social networks, so that the analysis of data is more accurate and more convincing, the contact relationship between the cows is reflected visually by the cow social networks, some cows with extremely low infection possibility are eliminated, and the analysis efficiency is improved; the epidemic risk level early warning is more accurate by combining the characteristics of each cow in multiple aspects and the characteristic information between contact cows, and the disease prevention and control are more timely and efficient.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a risk early warning method for epidemic prevention and control in animal husbandry according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a social network of a cow according to an embodiment of the present invention.
Detailed Description
In order to further explain the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description, the structure, the features and the effects of the risk pre-warning method for preventing and controlling epidemic diseases in animal husbandry according to the present invention are provided with the accompanying drawings and the preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The method is suitable for further performing prevention, control and early warning on all normal cows in a grassland after the sick cows with infectious diseases are found, and comprises the steps of firstly obtaining an attribute vector and a location vector of each cow, constructing a cow social network based on the contact condition among all cows, dividing the normal cows into dangerous cows and safe cows based on the contact condition among all cows in the cow social network, and respectively updating the attribute vectors and the location vectors corresponding to the dangerous cows and the safe cows; and analyzing based on the updated vector to obtain the epidemic disease risk level of the normal cow, and early warning workers according to different epidemic disease risk levels, so that the epidemic disease risk level is processed in a targeted manner, and the efficiency of the epidemic disease prevention and control process is improved.
The following describes a specific scheme of the risk early warning method for epidemic prevention and control in animal husbandry in detail with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a risk pre-warning method for epidemic prevention and control in animal husbandry according to an embodiment of the present invention is shown, where the method includes the following steps:
step S100, acquiring an attribute vector and a location vector of each cow, wherein the cows comprise normal cows and sick cows; and acquiring corresponding regional marginality and regional fixity based on the location vector of each cow.
Specifically, the method includes the steps that the cows in a grassland are monitored in real time, so that whether abnormal conditions exist in each cow, such as inappetence or lazy dispersion, can be found in time, when the cows are abnormal, relevant workers can check the cows to determine whether infectious diseases exist in the cows and the types of the infectious diseases, when the cows are detected to have the infectious diseases, the cows are marked as diseased cows, and whether other non-diseased normal cows are possibly infected or not is analyzed on the basis of the activity conditions of the diseased cows in a historical time period.
Marking all dairy cows, and distinguishing the diseased dairy cows from the normal dairy cows by marking different labels, but because all dairy cows are possibly contacted in the activity, each normal dairy cow is possibly infected by the disease; constructing an attribute vector corresponding to each cow based on the labeling information of each cow, wherein the elements in the attribute vector are as follows: label, age, breed and frequency of illness; digitally labeling each element in the attribute vector, for example, the label element value of the diseased cow is 1, and the label element value of the normal cow is 0; and the attribute vector corresponding to each cow is obtained by analogy.
Furthermore, a positioning device is arranged on the neck of each cow and used for monitoring the position of each cow in real time, in the embodiment of the invention, the position information of the cow is recorded and stored once every 0.1 hour, and the location vector corresponding to each cow is obtained; dividing a grassland area into a plurality of sub-areas, acquiring activity place information of each cow in a historical time period, constructing a corresponding place thermodynamic diagram according to the activity place information of each cow, and acquiring the heat of each sub-area according to the place thermodynamic diagrams; and selecting a plurality of sub-areas with the highest heat in all the sub-areas to sequence to obtain corresponding heat sequences, wherein the heat sequences are the corresponding location vectors of the cows.
The specific method for acquiring the location vector comprises the following steps: when the staff finds that the sick cows with infectious diseases appear in the grassland, all the normal cows which have contacted with the sick cows in the grassland need to be monitored in real time, the position data information of the sick cows in a historical period of time is traced, and an implementer in a specific period of time can determine the sick cows according to the actual morbidity latency.
Preferably, the time is set to be 8 days in the embodiment of the invention, namely, the position data information of the sick cow in the past 8 days is traced.
In order to more accurately and conveniently record the activity position information of each cow, the whole meadow area is divided into a plurality of sub-areas, each sub-area is labeled in sequence, the size of each sub-area can be set by an implementer, and the area of the circle at the outermost side of the meadow is labeled as an edge area.
Preferably, the size of the sub-regions divided into the grassland in the embodiment of the present invention is 1 square kilometer, that is, the area of each sub-region is 1 square kilometer.
The location thermodynamic diagrams corresponding to each cow are constructed according to the location data information of each cow in the past 8 days, the more dense the location and the higher the location heat, and the construction of the specific thermodynamic diagrams is a known technology and is not described any more. Constructing a corresponding location vector according to the location thermodynamic diagram corresponding to each cow, wherein the construction method of the location vector comprises the following steps: selecting a plurality of areas with the highest heat of the dairy cow in a site thermodynamic diagram, and arranging the areas according to the sequence of the heat from high to low to obtain a corresponding heat sequence, wherein the heat sequence is a site vector corresponding to the dairy cow.
Preferably, in the embodiment of the present invention, 32 regions with the highest heat are selected for arrangement, and when the milk cow has less than 32 regions in the past 8 days, 0 is supplemented to the location vector of the milk cow, so that the lengths of the location vectors corresponding to each milk cow are identical and are all 32 elements.
Further, acquiring the outermost sub-region in the grassland region as an edge region, counting the number of the edge regions in the corresponding location vector of each cow, wherein the ratio of the number of the edge regions to the number of all the sub-regions is the regional marginality of the cow; and acquiring the number of 0 element values in the location vector of the cow, and obtaining the regional fixity of the cow based on the number of 0 element values and the number of all elements in the location vector.
Considering that a sick cow may have an action of reducing social contact far away from the group or may be rejected by the group because of self-carried diseases, and therefore may be frequently located in the marginal area of the grassland, the marginal characteristic of the area corresponding to each cow is obtained based on the location vector corresponding to the cow:
Figure DEST_PATH_IMAGE009
wherein the content of the first and second substances,
Figure 402142DEST_PATH_IMAGE010
representing the marginality of the corresponding region of the cow;
Figure DEST_PATH_IMAGE011
representing the number of all edge areas in the corresponding location vector of the cow;
Figure 660255DEST_PATH_IMAGE012
representing the total number of sub-regions into which the grass field is divided.
Some cows may also have lassitude after illness, resulting in very lazy and not very sporty cows; some cows may run around due to pain of diseases, so that the diseased cows generally have a larger deviation in the number of arrival areas compared with normal cows, and therefore, the area fixity corresponding to each cow is used as an analysis index of the cow, and specifically, the area fixity corresponding to the cow is obtained based on the location vector as:
Figure 13876DEST_PATH_IMAGE014
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE015
representing the area fixity corresponding to the dairy cows;
Figure 467860DEST_PATH_IMAGE016
representing the number of 0 element values in the corresponding location vector of the cow;
Figure DEST_PATH_IMAGE017
the number of all elements in the corresponding location vector of the cow is shown, and the embodiment of the invention
Figure 371487DEST_PATH_IMAGE018
。
Step S200, connecting all the contact cows to construct a social network of the cows, wherein the contact is that the distance between any two cows is smaller than a preset range; and dividing all normal cows in the cow social network into dangerous cows and safe cows according to the connection relation in the cow social network.
Specifically, the position data information corresponding to each cow is obtained in step S100, so that the distance between the cows can be calculated according to the position data information corresponding to each cow, and when the distance between any two cows is smaller than a preset range, it indicates that there is contact between the two cows. Considering that the types of infectious diseases of the sick cows are different, and the corresponding infectious conditions are different, the embodiments of the invention divide the types of infectious diseases commonly suffered by cows into two main types, namely respiratory infectious diseases and skin infectious diseases; respiratory infectious diseases are generally transmitted through the digestive tract or the respiratory tract, so that for the sick cows with the respiratory infectious diseases, the preset range of the distance between the cows is set to be 1.6 meters, namely when the distance between any cow and the sick cow with the respiratory infectious diseases is less than 1.6 meters, the two cows are contacted; the cutaneous contagious diseases are generally transmitted by direct contact and indirect contact, so that for a diseased cow with cutaneous contagious diseases, a preset range of 0.8 m is set for the distance between cows, i.e. when the distance between any cow and the diseased cow with cutaneous contagious disease is less than 0.8 m, then there is contact between both cows.
Continuously monitoring position data information between two dairy cows in contact, and acquiring distance data corresponding to a plurality of moments to construct a distance sequence, wherein the distance data is acquired at intervals of every 0.1 hour in the embodiment of the invention; the contact degree between the cows is obtained based on the distance sequence as follows:
Figure 446891DEST_PATH_IMAGE020
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE021
denotes the first
Figure 17418DEST_PATH_IMAGE022
First cow and second cow
Figure DEST_PATH_IMAGE023
The degree of contact between the cows;
Figure 227207DEST_PATH_IMAGE024
denotes the first
Figure DEST_PATH_IMAGE025
The distance between two cows at each moment;
Figure 167481DEST_PATH_IMAGE026
indicating a predetermined range of distances, for a diseased cow suffering from a respiratory infectious disease,
Figure DEST_PATH_IMAGE027
in the case of a diseased cow with a cutaneous infectious disease,
Figure 877817DEST_PATH_IMAGE028
。
it should be noted that, the element values in the distance sequence are obtained by tracing the data of the two cows within 8 days; within 8 days, if the contact distance between two cows is greater than the preset range of the distance, the two cows do not have distance data therebetween, and therefore only data when the contact distance between the two cows is less than the preset range of the distance is included in the distance sequence.
Furthermore, each cow is used as a node, and when any two cows are contacted, the nodes corresponding to the two cows are connected; and connecting all the nodes corresponding to the contacted cows to obtain the social network of the cows. Acquiring a contact risk route corresponding to each diseased cow based on a social network of the cows, wherein the contact risk route is formed by all normal cows connected with the diseased cows, and the connection comprises direct connection and indirect connection; normal cows on the contact risk route are dangerous cows, and normal cows not on the contact risk route are safe cows. The sick dairy cows in the social network of the dairy cows are marked as 0 layer, the dangerous dairy cows directly connected with the sick dairy cows are 1 layer, the dangerous dairy cows directly connected with the dangerous dairy cows on the 1 layer are 2 layers, and the like, so that the layer number of the dangerous dairy cows in the social network of the dairy cows is marked.
Particularly, the dairy cow is an animal which likes social contact, and like all social animals, the dairy cow and the similar animals can feel safe and comfortable; within the same population, cows can often be closely related to another 2 to 4 cows; cows prefer to stay with friends and family for a sense of security if conditions permit. When the sick dairy cows exist in the grassland, other normal dairy cows contacting with the sick dairy cows are likely to be infected with the disease, so that each dairy cow is used as a node to construct a social network of the dairy cows, and the contact between the dairy cows in the embodiment of the invention refers to a preset range in which the distance between two dairy cows is smaller than a set distance; all dairy cows with contact within 8 days are obtained, corresponding nodes are connected when every two dairy cows are contacted, and by parity of reasoning, connection is carried out based on the contact condition between every two dairy cows to obtain a complete dairy cow social network.
Traversing by taking each diseased cow in the cow social network as a starting point to obtain all normal cows which are directly and indirectly connected with the diseased cow, so as to obtain a complete connection route, wherein the connection route is a contact risk route of the diseased cow; acquiring contact risk routes of all sick cows, wherein all normal cows on the contact risk routes are marked as dangerous cows; accordingly, normal cows that are not in the contact risk route in the cow social network are safe cows.
Stratifying normal cows of each node based on the diseased cows; each diseased cow is labeled as level 0 in the social network of that cow, and dangerous cows that are connected in direct contact with the diseased cow are labeled as level 1; marking dangerous cows which are in direct contact with the dangerous cows on the layer 1 as the layer 2; and by analogy, the number of layers corresponding to each dangerous cow in the cow social network is obtained.
It should be noted that there are no number of layers in the safe cows in the social network of cows and the safe cows in contact with the safe cows.
As a preferred example, please refer to fig. 2, which shows a schematic diagram of a social network of cows, wherein rectangular nodes represent sick cows and circular nodes represent normal cows; the contact risk route corresponding to the diseased cow H includes a normal cow f, a normal cow g, a normal cow e, a normal cow d, a normal cow b, a normal cow c, and a normal cow a; the contact risk routes corresponding to the diseased cow B comprise a normal cow a, a normal cow B, a normal cow c, a normal cow d, a normal cow e, a normal cow f and a normal cow g; the contact risk routes corresponding to the sick dairy cows N comprise normal dairy cows v, normal dairy cows x, normal dairy cows y and normal dairy cows z; therefore, in the dairy cow social network, normal dairy cow a, normal dairy cow b, normal dairy cow c, normal dairy cow d, normal dairy cow e, normal dairy cow f, normal dairy cow g, normal dairy cow v, normal dairy cow x, normal dairy cow y and normal dairy cow z are all dangerous dairy cows; and the normal cow h, the normal cow m and the normal cow r are all safe cows.
Wherein the diseased cow H, the diseased cow B and the diseased cow N are all 0 layers in the social network of the cows, the dangerous cow v directly connected with the diseased cow N is 1 layer, the dangerous cow x directly connected with the dangerous cow v is 2 layers, and the dangerous cow y and the dangerous cow z directly connected with the dangerous cow x are all 3 layers; for the dangerous cow f, when the dangerous cow f belongs to the contact risk route of the sick cow H, the dangerous cow f is 1 layer; when the dangerous dairy cow f belongs to the contact risk route of the sick dairy cow B, the dangerous dairy cow f is in 5 layers; by analogy, the number of layers corresponding to the normal cow a, the normal cow b, the normal cow c, the normal cow d, the normal cow e and the normal cow g respectively is obtained when the contact risk routes corresponding to different diseased cows are obtained.
Since the normal cow h, the normal cow m, and the normal cow r are safe cows, there is no number of layers.
Step S300, updating the location vector according to the regional marginality and the regional fixity corresponding to each normal cow in the cow social network to obtain a social environment vector; and updating the attribute vector of each dangerous cow in the cow social network to obtain a social attribute vector, and updating the attribute vector of each safe cow in the cow social network to obtain a social perception vector.
Specifically, in step S200, a dairy cow social network diagram corresponding to all dairy cows on the grassland is obtained, and the dairy cows corresponding to each node in the dairy cow social network diagram correspond to two vectors, which are an attribute vector and a location vector. In order to detect the health condition of each cow more accurately, the attribute vector and the location vector of each cow are updated based on the cow social network diagram.
(1) Firstly, updating nodes of normal cows on a contact risk route formed by each diseased cow, namely updating attribute vectors corresponding to dangerous cows on the contact risk route formed by each diseased cow; the updating sequence is from the dangerous cow on the layer 1 on the contact risk route to the dangerous cow on the highest layer on the contact risk route formed by the sick cow. Selecting a node corresponding to any dangerous cow in a cow social network as a point to be processed, acquiring all adjacent nodes directly connected with the point to be processed, acquiring a weight corresponding to each lower-layer neighborhood point, wherein the number of layers of all adjacent nodes is less than that of the point to be processed, and acquiring a social attribute vector of the point to be processed based on the weighted sum of the weights of all lower-layer neighborhood points and corresponding attribute vectors.
As an example, assume that the nth level of the contact risk route for any one sick cow is constructed
Figure 838820DEST_PATH_IMAGE022
The node of each dangerous cow is used as a point to be treated for updating, and the point to be treated is recorded as
Figure DEST_PATH_IMAGE029
And the point to be processed
Figure 535905DEST_PATH_IMAGE029
The node of all dangerous cows directly connected is the node
Figure 607766DEST_PATH_IMAGE029
The adjacent nodes of (1) are recorded as the points to be processed, the nodes which belong to the n-1 layer, the n-2 layer, the … layer and the n-n layer in all the adjacent nodes can exist in a plurality of layers of nodes
Figure 156297DEST_PATH_IMAGE029
The lower layer neighborhood point of the point to be processed
Figure 38933DEST_PATH_IMAGE029
The lower neighborhood set formed by all the lower neighborhood points is recorded as
Figure 564593DEST_PATH_IMAGE030
。
Based on the point to be processed
Figure 375615DEST_PATH_IMAGE029
Calculating the corresponding weight of each lower-layer neighborhood point by the corresponding attribute vector and the attribute vector of each lower-layer neighborhood point, wherein the specific algorithm of the weight is as follows:
Figure 686642DEST_PATH_IMAGE032
wherein, the first and the second end of the pipe are connected with each other,
Figure DEST_PATH_IMAGE033
represents the point to be processed
Figure 160087DEST_PATH_IMAGE029
Corresponding to the first in the lower neighborhood set
Figure 658195DEST_PATH_IMAGE023
The weight of each lower-layer neighborhood point;
Figure 337438DEST_PATH_IMAGE021
representing points to be processed
Figure 597912DEST_PATH_IMAGE029
And a first step of
Figure 87930DEST_PATH_IMAGE023
Degree of contact between individual sub-level neighbourhoods, i.e.
Figure 322602DEST_PATH_IMAGE034
First of a layer
Figure 789224DEST_PATH_IMAGE022
First dangerous cow and the first in the lower neighborhood set
Figure 933898DEST_PATH_IMAGE023
The degree of contact between the first dangerous cows;
Figure DEST_PATH_IMAGE035
representing points to be processed
Figure 893020DEST_PATH_IMAGE029
Corresponding attribute vectors, i.e.
Figure 614988DEST_PATH_IMAGE034
First of a layer
Figure 511400DEST_PATH_IMAGE022
Attribute vectors corresponding to the first dangerous cows;
Figure 353323DEST_PATH_IMAGE036
representing the first in the lower neighborhood set
Figure 44199DEST_PATH_IMAGE023
The attribute vector corresponding to each lower neighborhood point, i.e. the first in the lower neighborhood set
Figure 253463DEST_PATH_IMAGE023
Attribute vectors corresponding to the individual dangerous cows;
Figure DEST_PATH_IMAGE037
representing points to be processed
Figure 533659DEST_PATH_IMAGE029
All the lower layer neighborhood points form a lower layer neighborhood set;
Figure 761247DEST_PATH_IMAGE038
representing attribute vectors
Figure 13237DEST_PATH_IMAGE035
And attribute vector
Figure 460530DEST_PATH_IMAGE036
Cosine similarity therebetween;
Figure DEST_PATH_IMAGE039
a calculation representing an exponential function;
Figure 391753DEST_PATH_IMAGE040
the weight coefficient is set by the implementer.
Preferably, the embodiment of the invention is provided with
Figure DEST_PATH_IMAGE041
(ii) a Because each cow has own immunity, even if contact action exists between every two cows, the cows may not be infected with diseases due to own immunity, so the weight of the contact degree is properly reduced, and the follow-up score is dividedThe separation is more reasonable.
By parity of reasoning, the point to be processed is obtained
Figure 162263DEST_PATH_IMAGE029
The weight corresponding to all the lower layer neighborhood points is calculated based on the weighted sum between the weight of each lower layer neighborhood point and the attribute vector
Figure 834421DEST_PATH_IMAGE029
Updating the corresponding attribute vector, specifically:
Figure DEST_PATH_IMAGE043
wherein the content of the first and second substances,
Figure 565748DEST_PATH_IMAGE044
representing points to be processed
Figure 109118DEST_PATH_IMAGE029
An updated attribute vector;
Figure 62031DEST_PATH_IMAGE035
representing points to be processed
Figure 406556DEST_PATH_IMAGE029
An attribute vector before updating;
Figure 77708DEST_PATH_IMAGE033
represents the point to be processed
Figure 562785DEST_PATH_IMAGE029
Corresponding to the first in the lower neighborhood set
Figure 370204DEST_PATH_IMAGE023
The weight of each lower-layer neighborhood point;
Figure 151209DEST_PATH_IMAGE036
representing the first in the lower neighborhood set
Figure 575238DEST_PATH_IMAGE023
The attribute vector corresponding to each of the lower neighborhood points, i.e. the first in the set of lower neighborhoods
Figure 335777DEST_PATH_IMAGE023
Attribute vectors corresponding to the individual dangerous cows;
Figure 263281DEST_PATH_IMAGE037
representing points to be processed
Figure 480767DEST_PATH_IMAGE029
And all the lower neighborhood points form a lower neighborhood set.
And by analogy, obtaining the updated attribute vector corresponding to each dangerous cow on the contact risk route of any sick cow, recording the updated attribute vector as the social attribute vector of the dangerous cow, and making the attribute vectors corresponding to the lower neighborhood point in the updating process of the point to be processed be the updated attribute vectors, namely the social attribute vectors.
Based on the method for obtaining the same social attribute vector of the dangerous cow on the contact risk route of one diseased cow, the social attribute vectors of the dangerous cows on the contact risk routes of all the diseased cows are obtained, and the order of analyzing the contact risk routes of each diseased cow is carried out according to the time sequence for finding the diseased cow.
(2) Then, updating the attribute vector of each safe cow in the cow social network; selecting a node corresponding to any safe cow in the cow social network as a target point, obtaining a neighborhood node set according to all nodes directly connected with the target point, obtaining the weight of each node in the neighborhood node set, and obtaining the social perception vector of the target point based on the weighted summation of the weights of all nodes in the neighborhood node set and the corresponding attribute vectors.
As an example, assuming that any safe cow corresponding node in the cow social network is updated as a target point, the attribute vector of the safe cow corresponding to the target point is recorded as
Figure DEST_PATH_IMAGE045
All the safe dairy cow nodes directly connected with the target point are neighborhood nodes of the target point, and all the neighborhood nodes form a neighborhood node set
Figure 172517DEST_PATH_IMAGE046
(ii) a Obtaining the weight corresponding to any neighborhood node in the neighborhood node set as:
Figure 218971DEST_PATH_IMAGE048
wherein, the first and the second end of the pipe are connected with each other,
Figure DEST_PATH_IMAGE049
indicating correspondence of target points
Figure 190862DEST_PATH_IMAGE022
Neighbor node set of head safe cows
Figure 297359DEST_PATH_IMAGE023
The weight corresponding to the first safe cow;
Figure 915553DEST_PATH_IMAGE050
indicating a point correspondence of the target point
Figure 765697DEST_PATH_IMAGE022
First safe cow and the first in its neighborhood node set
Figure 917062DEST_PATH_IMAGE023
The degree of contact between safe-head cows;
Figure 69826DEST_PATH_IMAGE045
indicating a point correspondence of the target point
Figure 159004DEST_PATH_IMAGE022
An attribute vector of a safe-head cow;
Figure DEST_PATH_IMAGE051
representing the first in a set of neighborhood nodes
Figure 483280DEST_PATH_IMAGE023
An attribute vector of a safe-head cow;
Figure 974304DEST_PATH_IMAGE052
representing attribute vectors
Figure 671871DEST_PATH_IMAGE045
And attribute vector
Figure 779504DEST_PATH_IMAGE051
Cosine similarity therebetween;
Figure DEST_PATH_IMAGE053
representing a neighborhood node set corresponding to the target point;
Figure 784500DEST_PATH_IMAGE039
a calculation representing an exponential function;
Figure 506862DEST_PATH_IMAGE054
the weight coefficient is set by the implementer.
Preferably, the device is arranged in the embodiment of the invention
Figure DEST_PATH_IMAGE055
。
By analogy, the weight of the neighborhood node corresponding to each safe cow in the neighborhood node set corresponding to the target point is obtained, so that the updated attribute vector of the safe cow corresponding to the target point is obtained as follows:
Figure DEST_PATH_IMAGE057
wherein the content of the first and second substances,
Figure 860483DEST_PATH_IMAGE058
representCorresponding to the target point
Figure 35506DEST_PATH_IMAGE022
Updated attribute vectors for individual safe cows;
Figure 296723DEST_PATH_IMAGE045
indicating a point correspondence of the target point
Figure 247492DEST_PATH_IMAGE022
An attribute vector of each safe cow before updating;
Figure 568752DEST_PATH_IMAGE051
representing the first in a neighborhood node set
Figure 369087DEST_PATH_IMAGE023
An attribute vector of each safe cow;
Figure 168416DEST_PATH_IMAGE049
indicating correspondence of target points
Figure 504850DEST_PATH_IMAGE022
Neighbor node set of head safe cows
Figure 108263DEST_PATH_IMAGE023
The weight corresponding to the safe-head cow;
Figure 881047DEST_PATH_IMAGE053
and representing a neighborhood node set corresponding to the target point.
It should be noted that, when an updated neighbor node of a safe cow exists in a neighbor node set of a safe cow corresponding to a certain target point, weighted summation is performed based on the updated attribute vector of the safe cow in the neighbor node set.
And by analogy, obtaining the updated attribute vector corresponding to each safe cow in the cow social network, and recording the updated attribute vector of each safe cow as the social perception vector.
In order to improve the updating accuracy of each node, based on the method for obtaining the same social perception vector of each safe cow, the social perception vector of each safe cow in the cow social network is updated iteratively; in order to prevent an over-smoothing phenomenon in an iteration process and avoid that updated information of all nodes tends to be consistent due to excessive iteration times, an iteration stop condition is set in the embodiment of the present invention, which specifically includes:
Figure 234799DEST_PATH_IMAGE060
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE061
representing an iteration stop index;
Figure 721013DEST_PATH_IMAGE062
denotes the first
Figure 728284DEST_PATH_IMAGE022
The social perception vector corresponding to each safe cow during current iteration updating;
Figure DEST_PATH_IMAGE063
is shown as
Figure 556738DEST_PATH_IMAGE022
The social perception vector corresponding to each safe cow during the last iteration updating;
Figure 432290DEST_PATH_IMAGE064
representing the number of all safe cows in the social network of cows;
Figure DEST_PATH_IMAGE065
representing the modulus of the calculated vector.
And when the iteration stop index is smaller than a preset threshold value, stopping the update of the social perception vector for the safe dairy cow in the social network of the dairy cow.
Preferably, in the embodiment of the present invention, the preset threshold is set to 0.2, that is, when the value of the iteration stop index is less than 0.2, the update of the social perception vector of the safe cow in the social network of the cow is stopped.
Furthermore, whether each cow is infected with a disease is also closely related to the environment, and when a diseased cow contacts certain subregions on a grassland, a normal cow contacts the same subregions, so that the possibility of indirect infection exists.
(3) Finally, updating the node corresponding to each normal cow in the cow social network, namely updating the location vector corresponding to each normal cow in the cow social network; selecting nodes corresponding to any normal cow in the cow social network as points to be updated, obtaining a neighborhood node set according to all nodes directly connected with the points to be updated, calculating to obtain corresponding weights according to regional edge lines and regional fixity corresponding to each node in the neighborhood node set, and obtaining a social environment vector of the points to be updated based on weighted summation of the weights of all nodes in the neighborhood node set and corresponding location vectors.
As an example, assume that it is currently the first to update
Figure 648377DEST_PATH_IMAGE022
The node corresponding to each normal cow is a point to be updated, and the location vector corresponding to the point to be updated is recorded as
Figure 951182DEST_PATH_IMAGE066
All the nodes directly connected to the point to be updated are the neighborhood nodes of the point to be updated, and all the neighborhood nodes form a neighborhood node set
Figure 449291DEST_PATH_IMAGE046
(ii) a Obtaining the weight corresponding to each neighborhood node in the neighborhood node set as follows:
Figure 394113DEST_PATH_IMAGE068
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE069
indicates the second to be updated
Figure 467636DEST_PATH_IMAGE022
Neighbor node of a node set
Figure 472501DEST_PATH_IMAGE023
Weights corresponding to the neighborhood nodes;
Figure 956441DEST_PATH_IMAGE070
representing the first in a neighborhood node set
Figure 439375DEST_PATH_IMAGE023
The regional marginality corresponding to each neighborhood node;
Figure DEST_PATH_IMAGE071
representing the first in a neighborhood set of nodes
Figure 521731DEST_PATH_IMAGE023
The area fixity corresponding to each neighborhood node;
Figure 543170DEST_PATH_IMAGE066
indicates the second to be updated
Figure 15871DEST_PATH_IMAGE022
A location vector of each node;
Figure 36917DEST_PATH_IMAGE072
representing the first in a neighborhood set
Figure 675577DEST_PATH_IMAGE023
A location vector of each neighborhood node;
Figure 756666DEST_PATH_IMAGE039
a calculation representing an exponential function;
Figure DEST_PATH_IMAGE073
showing groundPoint vector
Figure 155811DEST_PATH_IMAGE066
And location vector
Figure 980547DEST_PATH_IMAGE072
Cosine similarity between them;
Figure 709600DEST_PATH_IMAGE053
representing a neighborhood node set;
Figure 227169DEST_PATH_IMAGE074
a weight coefficient representing the marginal property of the region, which is set by an implementer;
Figure DEST_PATH_IMAGE075
the weight coefficient representing the region fixing degree is set by the implementer.
Preferably, the embodiments of the present invention are empirically set
Figure 969735DEST_PATH_IMAGE076
，
Figure DEST_PATH_IMAGE077
。
By parity of reasoning, the current the second one to be updated is obtained
Figure 268603DEST_PATH_IMAGE022
The weight of each neighborhood node in the neighborhood node set of each node is weighted and summed based on the weights corresponding to all the neighborhood nodes in the neighborhood node set to obtain the second
Figure 367009DEST_PATH_IMAGE022
The location vector after the update of each node is as follows:
Figure DEST_PATH_IMAGE079
wherein the content of the first and second substances,
Figure 602949DEST_PATH_IMAGE080
indicates the second to be updated
Figure 36074DEST_PATH_IMAGE022
The location vector after the update of each node;
Figure 812400DEST_PATH_IMAGE066
indicates the second to be updated
Figure 30891DEST_PATH_IMAGE022
A location vector before updating of each node;
Figure 876881DEST_PATH_IMAGE072
indicates the second to be updated
Figure 548034DEST_PATH_IMAGE022
Neighbor node set of individual node
Figure 268996DEST_PATH_IMAGE023
A location vector of each neighborhood node;
Figure 76415DEST_PATH_IMAGE069
indicates the second to be updated
Figure 355956DEST_PATH_IMAGE022
Neighbor node of a node set
Figure 858612DEST_PATH_IMAGE023
Weights corresponding to the neighborhood nodes;
Figure 101375DEST_PATH_IMAGE053
representing a set of neighborhood nodes.
By analogy, obtaining the updated location vector of the node corresponding to each normal cow in the cow social network, and recording the updated location vector of each normal cow as a social environment vector; when it is to be updated
Figure 140131DEST_PATH_IMAGE022
And when the updated neighborhood nodes exist in the neighborhood node set of each node, performing weighted summation calculation by using the social environment vector which is favorable for updating the nodes.
Further, after all nodes corresponding to all normal cows in the social network of the cows are updated, iterative updating is continued, so as to avoid the phenomenon of over-smoothing, an early stop method is adopted in the embodiment of the invention, and the calculation environment iteration stop index is as follows:
Figure 826459DEST_PATH_IMAGE082
wherein the content of the first and second substances,
Figure 941045DEST_PATH_IMAGE061
representing an environment iteration stop index;
Figure DEST_PATH_IMAGE083
denotes the first
Figure 908870DEST_PATH_IMAGE022
The social environment vector corresponding to each normal cow during current iteration updating;
Figure 690881DEST_PATH_IMAGE084
denotes the first
Figure 49575DEST_PATH_IMAGE022
The corresponding social environment vector of each normal cow during the last iteration updating;
Figure DEST_PATH_IMAGE085
representing the number of all normal cows in the cow social network;
Figure 323561DEST_PATH_IMAGE065
representing the modulus of the calculated vector.
And when the environment iteration stop index is smaller than a preset threshold value, stopping updating the social environment vector for all normal cows in the cow social network.
Preferably, in the embodiment of the present invention, the preset threshold is set to 0.1, that is, when the value of the environment iteration stop index is less than 0.1, the update of the social environment vectors of all normal cows in the social network of the cow is stopped.
Step S400, obtaining a target attribute vector and a target location vector of the sick cow, and calculating attribute similarity between each normal cow and the sick cow as a first similarity, wherein the attribute similarity comprises cosine similarity between a social attribute vector and the target attribute vector and cosine similarity between a social perception vector and the target attribute vector; and calculating cosine similarity between the social environment vector and the target location vector of each normal cow to be second similarity, obtaining the epidemic risk level of the normal cow based on the first similarity and the second similarity of each normal cow, and timely early warning based on the epidemic risk level.
Obtaining a social attribute vector and a social environment vector corresponding to each dangerous cow in the cow social network, and a social perception vector and a social environment vector corresponding to each safe cow in the step S300; and calculating the epidemic risk level corresponding to each normal cow based on two vectors corresponding to each normal cow in the social network of the cows.
Specifically, a normal cow in the social network of the cow is arbitrarily selected, and assuming that the normal cow is a dangerous cow, the method for calculating the epidemic risk level of the normal cow comprises the following steps:
firstly, acquiring attribute vectors and location vectors corresponding to all sick cows in a cow social network, calculating the mean value of the attribute vectors corresponding to all the sick cows as a target attribute vector, and calculating the mean value of the location vectors corresponding to all the sick cows as a target location vector; and taking the target attribute vector and the target location vector as the feature vector of each sick cow.
Then, calculating a first similarity of the attributes of the normal cow and the diseased cow and a second similarity of the environment of the normal cow and the diseased cow, namely calculating the cosine similarity between the social attribute vector of the normal cow and the target attribute vector of the diseased cow as the first similarity; and calculating the cosine similarity between the social environment vector of the normal cow and the target environment vector of the sick cow to be a second similarity.
And finally, obtaining corresponding epidemic risk grades based on the first similarity and the second similarity of the normal cows as follows:
Figure 422973DEST_PATH_IMAGE086
wherein, the first and the second end of the pipe are connected with each other,
Figure 669278DEST_PATH_IMAGE003
representing the epidemic risk level corresponding to the normal cow;
Figure 946676DEST_PATH_IMAGE004
representing a first similarity of the normal cow;
Figure 682858DEST_PATH_IMAGE005
a second similarity representing the normal cow;
Figure 71114DEST_PATH_IMAGE006
represents an adjustment coefficient;
Figure 936039DEST_PATH_IMAGE007
indicating rounding up.
Preferably, the adjusting coefficient is set in the embodiment of the invention
Figure DEST_PATH_IMAGE087
。
And in the same way, acquiring epidemic risk levels corresponding to all normal cows in the social network of the cows.
It should be noted that, when the epidemic risk level calculation is performed on a normal cow, if the normal cow is a safe cow in the cow social network, the first similarity of the normal cow is obtained from the cosine similarity between the social perception vector and the target attribute vector corresponding to the normal cow.
Because the value ranges of the first similarity and the second similarity in the embodiment of the invention are both
Figure 666229DEST_PATH_IMAGE088
Therefore, the acquired epidemic disease risk level of each normal cow is between 0 and 10, and the early warning treatment is performed on the grassland staff in time based on the epidemic disease risk level corresponding to each normal cow, so that the staff can perform different treatment measures on the normal cows with different epidemic disease risk levels, for example, the normal cows with the epidemic disease risk level not less than 6 can be subjected to physical examination in time, and the normal cows with the epidemic disease risk level less than 6 can be subjected to observation or isolation and other measures, so that the large-area outbreak of the epidemic disease can be prevented.
In summary, the embodiment of the present invention provides a risk early warning method for epidemic prevention and control in animal husbandry, which includes obtaining an attribute vector and a location vector of each cow for analysis, constructing a cow social network based on contact conditions between a diseased cow and a normal cow in all cows, dividing the normal cows in the cow social network into dangerous cows and safe cows according to a connection relationship in the cow social network, and updating the attribute vectors corresponding to the dangerous cows and the safe cows to obtain corresponding social attribute vectors and social perception vectors; updating the location vectors of all normal cows in the cow social network to obtain corresponding social environment vectors; the method comprises the steps of further obtaining target environment vectors and target attribute vectors corresponding to all the sick cows, obtaining a first similarity corresponding to each normal cow based on the target attribute vectors, obtaining a second similarity corresponding to each normal cow based on the target environment vectors, obtaining epidemic risk grades of the normal cows by combining the first similarity and the second similarity, and early warning workers based on different epidemic risk grades, so that the treatment is performed pertinently, and the efficiency in the epidemic prevention and control process is improved.
It should be noted that: the sequence of the above embodiments of the present invention is only for description, and does not represent the advantages or disadvantages of the embodiments. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and should not be taken as limiting the scope of the present invention, which is intended to cover any modifications, equivalents, improvements, etc. within the spirit of the present invention.

Claims (10)

1. A risk early warning method for epidemic prevention and control in animal husbandry is characterized by comprising the following steps:
acquiring an attribute vector and a location vector of each cow, wherein the cows comprise normal cows and sick cows; acquiring corresponding regional marginality and regional fixity based on the location vector of each cow;
connecting all dairy cows with contact to construct a dairy cow social network, wherein the contact is that the distance between any two dairy cows is smaller than a preset range; dividing all normal cows in the cow social network into dangerous cows and safe cows according to the connection relation in the cow social network;
updating the location vector according to the regional marginality and the regional stationarity corresponding to each normal cow in the social network of the cows to obtain a social environment vector; updating the attribute vector of each dangerous cow in the cow social network to obtain a social attribute vector, and updating the attribute vector of each safe cow in the cow social network to obtain a social perception vector;
acquiring a target attribute vector and a target location vector of the sick dairy cows, and calculating attribute similarity between each normal dairy cow and the sick dairy cows as a first similarity, wherein the attribute similarity comprises cosine similarity between a social attribute vector and the target attribute vector and cosine similarity between a social perception vector and the target attribute vector; calculating cosine similarity between the social environment vector of each normal cow and the target location vector to be second similarity, obtaining the epidemic risk level of each normal cow based on the first similarity and the second similarity of each normal cow, and timely early warning based on the epidemic risk level.
2. The risk pre-warning method for controlling epidemic disease in animal husbandry according to claim 1, wherein the step of classifying all normal cows in the social network of cows into dangerous cows and safe cows according to the connection relationship in the social network of cows comprises:
acquiring a contact risk route corresponding to each diseased cow based on the social network of the cows, wherein the contact risk route is composed of all normal cows connected with the diseased cows, and the connection comprises direct connection and indirect connection; normal cows on the contact risk route are dangerous cows, and normal cows not on the contact risk route are safe cows.
3. The risk early warning method for epidemic prevention and control in animal husbandry according to claim 1, wherein the step of connecting all the contacted cows to construct a social network of cows comprises:
taking each cow as a node, and connecting the nodes corresponding to two cows when any two cows are contacted; and connecting all the nodes corresponding to the contacted cows to obtain a social network of the cows.
4. The risk early warning method for epidemic prevention and control in animal husbandry according to claim 3, wherein the step of updating the location vector according to regional marginality and regional stationarity corresponding to each normal cow in the social network of the cows to obtain a social environment vector comprises:
selecting nodes corresponding to any normal cow in the cow social network as points to be updated, obtaining a neighborhood node set according to all nodes directly connected with the points to be updated, calculating to obtain corresponding weights according to regional edge lines and regional fixity corresponding to each node in the neighborhood node set, and obtaining a social environment vector of the points to be updated based on the weighted summation of the weights of all nodes in the neighborhood node set and corresponding location vectors.
5. The risk early warning method for epidemic prevention and control in animal husbandry according to claim 3, wherein the step of updating the attribute vector of each dangerous cow in the social network of cows to obtain a social attribute vector comprises:
marking the diseased cows in the social network of the cows as 0 layer, the dangerous cows directly connected with the diseased cows as 1 layer, the dangerous cows directly connected with the dangerous cows on the 1 layer as 2 layers, and so on, and marking the number of layers of dangerous cows in the social network of the cows;
selecting a node corresponding to any dangerous cow in the cow social network as a point to be processed, acquiring all adjacent nodes directly connected with the point to be processed, wherein the number of layers in all adjacent nodes is smaller than that of the point to be processed, the adjacent nodes are lower-layer neighborhood points, acquiring a weight corresponding to each lower-layer neighborhood point, and acquiring a social attribute vector of the point to be processed based on weighted summation between the weights of all the lower-layer neighborhood points and corresponding attribute vectors.
6. The risk pre-warning method for epidemic prevention and control in animal husbandry according to claim 3, wherein the step of updating the attribute vector of each safe cow in the social network of cows to obtain the social perception vector comprises:
selecting a node corresponding to any safe cow in the cow social network as a target point, obtaining a neighborhood node set according to all nodes directly connected with the target point, obtaining the weight of each node in the neighborhood node set, and obtaining the social perception vector of the target point based on the weighted summation of the weights of all nodes in the neighborhood node set and corresponding attribute vectors.
7. The risk pre-warning method for epidemic prevention and control in animal husbandry according to claim 1, wherein the step of obtaining the epidemic risk rating of each normal cow based on the first similarity and the second similarity of each normal cow comprises:
the epidemic risk levels are:
Figure DEST_PATH_IMAGE001
wherein the content of the first and second substances,
Figure 596952DEST_PATH_IMAGE002
representing the epidemic risk level corresponding to the normal cow;
Figure 569325DEST_PATH_IMAGE003
representing a first similarity of the normal cow;
Figure 321380DEST_PATH_IMAGE004
representing a second similarity of the normal cow;
Figure 737318DEST_PATH_IMAGE005
represents an adjustment coefficient;
Figure 82980DEST_PATH_IMAGE006
indicating rounding up.
8. The risk early warning method for epidemic prevention and control in animal husbandry according to claim 1, wherein the step of obtaining the attribute vector of each cow comprises:
labeling the cow as a diseased cow or a normal cow as a label of the cow, elements in an attribute vector of the cow including: label, age, breed, and number of diseases.
9. The risk pre-warning method for epidemic prevention and control in animal husbandry according to claim 1, wherein the step of obtaining the location vector of each cow comprises:
dividing a grassland area into a plurality of sub-areas, acquiring activity place information of each cow in a historical time period, constructing a corresponding place thermodynamic diagram according to the activity place information of each cow, and acquiring the heat of each sub-area according to the place thermodynamic diagram;
and selecting a plurality of sub-areas with the highest heat in all the sub-areas to sequence to obtain corresponding heat sequences, wherein the heat sequences are the place vectors corresponding to the cows.
10. The risk early warning method for epidemic prevention and control in animal husbandry according to claim 9, wherein the step of obtaining corresponding regional marginality and regional stationarity based on the location vector of each cow comprises:
acquiring the outermost sub-region in the grassland region as an edge region, and counting the number of the edge regions in the location vector corresponding to each cow, wherein the ratio of the number of the edge regions to the number of all the sub-regions is the regional marginality of the cow;
and acquiring the number of 0 element values in the location vector of the cow, and obtaining the regional fixity of the cow based on the number of 0 element values and the number of all elements in the location vector.
CN202210714501.0A 2022-06-23 2022-06-23 Risk early warning method for epidemic disease prevention and control in animal husbandry Active CN114781952B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210714501.0A CN114781952B (en) 2022-06-23 2022-06-23 Risk early warning method for epidemic disease prevention and control in animal husbandry

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210714501.0A CN114781952B (en) 2022-06-23 2022-06-23 Risk early warning method for epidemic disease prevention and control in animal husbandry

Publications (2)

Publication Number Publication Date
CN114781952A true CN114781952A (en) 2022-07-22
CN114781952B CN114781952B (en) 2022-09-02

Family

ID=82422251

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210714501.0A Active CN114781952B (en) 2022-06-23 2022-06-23 Risk early warning method for epidemic disease prevention and control in animal husbandry

Country Status (1)

Country Link
CN (1) CN114781952B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107239882A (en) * 2017-05-10 2017-10-10 平安科技（深圳）有限公司 Methods of risk assessment, device, computer equipment and storage medium
US20180211718A1 (en) * 2014-12-24 2018-07-26 Stephan HEATH Systems, computer media, and methods for using electromagnetic frequency (emf) identification (id) devices for monitoring, collection, analysis, use and tracking of personal data, biometric data, medical data, transaction data, electronic payment data, and location data for one or more end user, pet, livestock, dairy cows, cattle or other animals, including use of unmanned surveillance vehicles, satellites or hand-held devices
CN112786210A (en) * 2021-01-15 2021-05-11 华南师范大学 Epidemic propagation tracking method and system
CN113938503A (en) * 2021-09-26 2022-01-14 云南追溯科技有限公司 Early warning system for diseases through live pig behavior sign monitoring and construction method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180211718A1 (en) * 2014-12-24 2018-07-26 Stephan HEATH Systems, computer media, and methods for using electromagnetic frequency (emf) identification (id) devices for monitoring, collection, analysis, use and tracking of personal data, biometric data, medical data, transaction data, electronic payment data, and location data for one or more end user, pet, livestock, dairy cows, cattle or other animals, including use of unmanned surveillance vehicles, satellites or hand-held devices
CN107239882A (en) * 2017-05-10 2017-10-10 平安科技（深圳）有限公司 Methods of risk assessment, device, computer equipment and storage medium
CN112786210A (en) * 2021-01-15 2021-05-11 华南师范大学 Epidemic propagation tracking method and system
CN113938503A (en) * 2021-09-26 2022-01-14 云南追溯科技有限公司 Early warning system for diseases through live pig behavior sign monitoring and construction method

Also Published As

Publication number Publication date
CN114781952B (en) 2022-09-02

Similar Documents

Publication Publication Date Title
US8297231B2 (en) System and methods for health monitoring of anonymous animals in livestock groups
US20100198024A1 (en) Vitality meter for health monitoring of anonymous animals in livestock groups
Eckelkamp et al. On-farm use of disease alerts generated by precision dairy technology
Steensels et al. A decision-tree model to detect post-calving diseases based on rumination, activity, milk yield, BW and voluntary visits to the milking robot
Firk et al. Improving oestrus detection by combination of activity measurements with information about previous oestrus cases
Maselyne et al. Online warning systems for individual fattening pigs based on their feeding pattern
KR102165891B1 (en) Livestock data analysis-based farm health state diagnosis system
Feiyang et al. Monitoring behavior of poultry based on RFID radio frequency network
Alshehri Blockchain-assisted internet of things framework in smart livestock farming
Kayser et al. Evaluation of statistical process control procedures to monitor feeding behavior patterns and detect onset of bovine respiratory disease in growing bulls
Cantor et al. Using machine learning and behavioral patterns observed by automated feeders and accelerometers for the early indication of clinical bovine respiratory disease status in preweaned dairy calves
Lawrence et al. Prevalence of failure of passive transfer of maternal antibodies in dairy calves in the Manawatu region of New Zealand
Bausewein et al. Sensitivity and specificity for the detection of clinical mastitis by automatic milking systems in bavarian dairy herds
Lupo et al. Bayesian network as an aid for Food Chain Information use for meat inspection
KR102274263B1 (en) Livestock health status immune standardization system
CN114781952B (en) Risk early warning method for epidemic disease prevention and control in animal husbandry
Steeneveld et al. Simplify the interpretation of alert lists for clinical mastitis in automatic milking systems
Mikail et al. Subclinical mastitis prediction in dairy cattle by application of Fuzzy Logic.
Exadaktylos et al. Chapter Automatic Identification and Interpretation of Animal Sounds, Application to Livestock Production Optimisation
Kayser et al. Efficacy of statistical process control procedures to identify deviations in continuously measured physiologic and behavioral variables in beef steers experimentally challenged with Mannheimia haemolytica
Lupo et al. Feasibility of screening broiler chicken flocks for risk markers as an aid for meat inspection
CN113989745A (en) Non-contact monitoring method for feeding condition of ruminants
Mate et al. Design and development of IoT-based intelligent solutions with blockchain for indian farmers on livestock management
De Waele et al. Internet of animals: unsupervised foaling detection based on accelerometer data
JPWO2021033732A5 (en)

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant