CN118014664A - User portrait construction method and alarm threshold self-adaptive setting method - Google Patents
User portrait construction method and alarm threshold self-adaptive setting method Download PDFInfo
- Publication number
- CN118014664A CN118014664A CN202410411821.8A CN202410411821A CN118014664A CN 118014664 A CN118014664 A CN 118014664A CN 202410411821 A CN202410411821 A CN 202410411821A CN 118014664 A CN118014664 A CN 118014664A
- Authority
- CN
- China
- Prior art keywords
- gas
- user
- service data
- tag
- gas consumption
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000010276 construction Methods 0.000 title claims abstract description 17
- 230000008859 change Effects 0.000 claims abstract description 6
- 238000003062 neural network model Methods 0.000 claims description 11
- 230000003044 adaptive effect Effects 0.000 claims description 6
- 238000003064 k means clustering Methods 0.000 claims description 5
- 238000004519 manufacturing process Methods 0.000 claims description 4
- 238000007689 inspection Methods 0.000 claims description 3
- 238000012549 training Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 238000004458 analytical method Methods 0.000 abstract description 5
- 239000007789 gas Substances 0.000 description 241
- 230000006399 behavior Effects 0.000 description 25
- 238000010586 diagram Methods 0.000 description 7
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 6
- 238000012795 verification Methods 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 230000035945 sensitivity Effects 0.000 description 3
- 239000007795 chemical reaction product Substances 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010411 cooking Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000003287 bathing Methods 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a construction method of a user image, which comprises the following steps: acquiring gas service data, wherein the gas service data comprises controlled gas service data and uncontrolled gas service data; identifying the gas service data through a pre-established label model to generate a user label; summarizing the user labels to obtain user portraits; and automatically updating the user portrait according to the change of the gas service data. According to the user portrait construction method provided by the embodiment of the application, the controlled gas service data and the uncontrolled gas service data are combined, and the generated user portrait is representative to the user, so that the analysis of the gas user can be better realized, and more accurate and personalized service is provided for the user. The application also provides a self-adaptive setting method of the alarm threshold, which is used for self-adaptive adjustment of the alarm threshold based on the user portrait, so that the alarm threshold can adapt to different user groups.
Description
Technical Field
The application relates to the technical field of gas business, in particular to a user portrait construction method and an alarm threshold self-adaptive setting method.
Background
For the gas business field, more accurate and personalized services are typically provided to users by collecting and analyzing their controlled data. Controlled data, among other things, refers to data that is stored, processed, and transmitted under strict management and control. Such controlled data originates from internal controlled systems (e.g., hour usage data, usage details, etc.), and generally complies with data governance policies and standards within the organization, including requirements in terms of data quality, security, privacy protection, and compliance.
However, the gas users have a plurality of groups, and different users have different types, numbers and service lives of gas using equipment owned by different users, and the information can influence the gas using behaviors of the users but cannot be acquired by an internal controlled system. Thus, users cannot be well analyzed based solely on the controlled data monitored by the internal controlled system.
Disclosure of Invention
The application aims to provide a technical scheme for better analyzing the gas users, so that more accurate and personalized services can be provided for the users.
Based on the above objects, the embodiment of the present application provides a method for constructing a user image, the method comprising the steps of:
Acquiring gas service data, wherein the gas service data comprises controlled gas service data and uncontrolled gas service data, and the controlled gas service data is derived from an internal controlled system data source of a service company; the uncontrolled gas service data is derived from a data source other than the internal controlled system data source;
Identifying the gas service data through a pre-established label model to generate a user label;
summarizing the user labels to obtain user portraits;
and automatically updating the user portrait according to the change of the gas service data.
Further, the tag model builds features based on the gas business data, the features are used for characterizing gas users, and the tag model identifies the features to generate user tags;
The user tag includes:
The gas consumption device comprises a user behavior characteristic label, a gas consumption characteristic label and a gas consumption equipment characteristic label, wherein the user behavior characteristic label is used for reflecting basic attributes of a gas user; the gas consumption characteristic tag is used for reflecting the gas consumption attribute of the gas user, and the gas consumption equipment characteristic tag is used for reflecting the gas consumption equipment attribute of the gas user;
The user behavior characteristic label, the gas utilization characteristic label and the gas utilization equipment characteristic label are directly obtained through gas service data or are obtained through gas service data inference.
Further, the basic attributes of the gas user include: lifestyle attributes, family structure attributes, economic capability attributes, and operational behavior attributes;
the gas consumption attributes of the gas user include: gas utilization behavior attributes and production operation attributes;
The gas utilization equipment attributes include: a gas usage type attribute and a gas usage parameter attribute.
Further, the construction method further comprises the following steps:
Based on the user labels, the user groups are clustered through an improved K-means clustering algorithm, different user clusters are constructed, comprehensive images of the users are formed, and visual presentation of the results is completed.
Further, the internally controlled system data source includes: an indoor gas utilization terminal, a work order security inspection system, a customer service and income system, a call telephone traffic system and a mobile client;
Uncontrolled gas business data sources include: social media, published network resources, third party data providers, and databases within business companies that are not managed and audited.
Further, uncontrolled gas service data actively uploaded by the gas user is received through the user uploading interface so as to update the user portrait.
Further, the method further comprises:
uncontrolled gas business data is periodically collected to update the user profile.
The application also provides a self-adaptive setting method of the alarm threshold, which comprises the following steps:
acquiring user portraits of a plurality of gas users, wherein the user portraits are generated based on the user portraits constructing method;
based on the user portrait, clustering the gas users to obtain different user clusters;
and setting an alarm threshold according to each cluster of gas users.
Further, the user portraits comprise a user behavior characteristic label, a gas utilization characteristic label and a gas utilization equipment characteristic label,
The user behavior feature tag is used for reflecting the basic attribute of the gas user; the gas consumption characteristic tag is used for reflecting the gas consumption attribute of the gas user, and the gas consumption equipment characteristic tag is used for reflecting the gas consumption equipment attribute of the gas user;
the method may further comprise the steps of,
Based on the user portrait of each user cluster, acquiring a user behavior characteristic label and an air consumption equipment characteristic label of each user cluster;
based on the user behavior feature tag and the gas utilization equipment feature tag, extracting the basic attribute and the gas utilization equipment attribute of the gas user;
And setting an initial value of an alarm threshold value based on the basic attribute and the gas using equipment attribute and combining the maximum load flow of the gas using equipment.
Further, historical gas consumption in the gas business data is obtained, the historical gas consumption is sequenced according to time sequence to obtain a gas consumption time sequence, and fast Fourier transform is carried out on the gas consumption time sequence;
calculating the amplitude of each frequency component, and finding out the frequency with the maximum amplitude after obtaining the amplitude of each frequency;
and converting the maximum frequency into a period length to obtain an estimated value of the threshold updating period.
Further, an LSTM neural network model is established, and training is carried out on the LSTM neural network model based on the characteristic labels of the user portraits;
predicting the gas consumption trend based on the trained LSTM neural network model;
based on the gas consumption trend and a preset safety standard, judging whether the gas user has potential safety risk,
If the potential safety risk is judged, the alarm threshold value is reduced; and if no safety risk is judged, the alarm threshold value is improved.
According to the description, the user portrait construction method provided by the embodiment of the application combines the controlled gas service data and the uncontrolled gas service data, and the generated user portrait is representative to the user, so that the analysis of the gas user can be better realized, and more accurate and personalized service is provided for the user.
Drawings
FIG. 1 is a flowchart of a user portrait construction method provided by an embodiment of the present application;
FIG. 2 is a schematic diagram of a first user portrait provided by an embodiment of the present application;
FIG. 3 is a schematic diagram of a second user representation provided by an embodiment of the present application;
FIG. 4 is a flowchart of an alarm threshold adaptive setting method according to an embodiment of the present application;
FIG. 5 is a schematic diagram of a gas consumption time sequence according to an embodiment of the present application;
Fig. 6 is a schematic diagram of the spectrum obtained after fourier transform of fig. 5.
Detailed Description
The present application will be described in detail below with reference to the specific embodiments shown in the drawings, but these embodiments are not limited to the present application, and structural, method, or functional modifications made by those skilled in the art based on these embodiments are included in the scope of the present application.
As shown in FIG. 1, an embodiment of the present application provides a user portrait construction method, which includes the following steps:
Step S11, acquiring gas service data, wherein the gas service data comprises controlled gas service data and uncontrolled gas service data;
step S12, marking the gas service data through a pre-established label model to generate a user label;
s13, summarizing the user labels to obtain user portraits;
And S14, automatically updating the user portrait according to the change of the gas service data.
Wherein the controlled gas service data is derived from an internal controlled system data source of the service company, the internal controlled system data source comprising: indoor gas utilization terminals, worksheet security inspection systems, customer service and revenue systems, call traffic systems, mobile clients and the like. In the embodiment of the application, the data which comes from the outside of the data source of the internal controlled system is called uncontrolled gas business data. Uncontrolled gas service data is typically not under strict regulatory control, and the source of uncontrolled gas service data may come from external sources, such as social media, public network resources, third party data providers, etc., or from databases within the service company that are not formally managed and audited.
The source of the uncontrolled gas service data is various, so that after the controlled and uncontrolled gas service data are combined, the obtained gas service data have the characteristics of multiple dimensions and fine granularity, and the image of a user is improved based on the controlled and uncontrolled combined gas service data, so that the analysis of the user is facilitated.
As an alternative implementation, the tag model builds features based on the gas business data, the features being used to symbolize gas users, the tag model identifying the features to generate user tags. The user tag includes: user behavior characteristic labels, gas consumption characteristic labels and gas consumption equipment characteristic labels;
The user behavior characteristic label, the gas utilization characteristic label and the gas utilization equipment characteristic label are directly obtained through gas service data or are obtained through gas service data inference.
As an alternative implementation, the user behavior feature tag is used to reflect basic attributes of the gas user, including life habit attributes (e.g., cooking habits, bathing habits, etc.), family structure attribute features (e.g., age, income, price of a user, whether at home for a long period of time, occupation of the user, etc.), economic capability attributes (e.g., resident consumption records, payment conditions, gas price sensitivity of an enterprise, etc.), and operational behaviors (e.g., operation habits of operation and maintenance personnel, maintenance frequency, emergency response capability, etc.). These underlying attributes may be embodied as behavioral characteristic tags of the user.
The gas consumption characteristic feature tag is used for reflecting gas consumption attributes of gas users, and relates to gas consumption behavior attributes (such as average gas consumption in month, fluctuation of gas consumption in month, tendency of gas theft, gas consumption scale of enterprises, gas consumption stability and the like) and production operation attributes (such as production process of enterprises, operation time and the like) of the users. These gas usage attributes may be embodied as gas usage characteristic feature tags for the user.
The gas utilization device feature tag is used for reflecting gas utilization device attributes of a gas user, wherein the gas utilization device attributes comprise gas utilization device type attributes and gas utilization device parameter attributes. For example, gas usage equipment type attributes include whether a residential user uses a gas water heater, whether a gas warmer is used, and a large gas usage equipment of an enterprise user; the gas consumption equipment parameter attributes comprise network environment conditions (such as 5g popularizing rate), security check service requirements, state of pipe network equipment, maintenance records, failure rate and the like. These gas usage device attributes may be embodied as a user's gas usage device feature tag.
In order to describe the process of identifying the gas service data and generating the user tag in step S12 in the user portrait construction method provided by the embodiment of the present application, a specific example will be described below.
It will be readily appreciated that some types of tags may be obtained directly from the gas service data, for example, for security check requirement tags, directly from controlled gas service data in the gas service data. The controlled gas service data includes battery power data collected by a controlled system within the service company. By reading the battery power data, whether the user has security check requirements or not can be directly known according to the battery power data, and a corresponding label is generated. In addition, the characteristic tags of gas consumption characteristics such as average gas consumption, fluctuation of gas consumption, tendency of gas theft or the like can also be obtained directly through gas business data. It should be noted that, the tag of the characteristic tag of using gas features, which has a tendency of stealing gas, can be obtained by deducing some uncontrolled gas service data besides being directly judged by the controlled gas service data. For example, when the valve shell is disassembled to give an alarm and electromagnetic interference to give an alarm, the suspicion that the user has gas stealing and leakage can be judged, corresponding labels are generated, and the user is listed in a gas integrity file.
Some types of tags are not directly available through controlled gas service data, but can be directly available through uncontrolled gas service data. For example, the age, occupation, home status, kitchen range status, etc. of the user may be collected by means of online reporting.
As an alternative implementation, for some tags obtained directly from uncontrolled gas service data, verification can also be inferred from the controlled gas service data.
Firstly, a controlled gas service data inference verification mode of some types of tags in user behavior feature tags is described:
For example, for a user age tag, the daily gas consumption time of the user is obtained through the controlled gas service data, the user is sketched, and if the user is useful gas in any period of the daytime, the user can be inferred to be a retired elderly user, daily working is not needed, and if the gas consumption has a time-period rule, the user can be inferred to be a daily regular work user. And comparing the user age label obtained by deducing the controlled gas service data with the user age information collected by the uncontrolled gas service data, so as to judge whether the uncontrolled gas service data is uploaded accurately.
And the income level label and the age label of the user can be deduced through the payment amount and the payment mode information in the controlled gas service data, the deduced result is compared with the uncontrolled gas service data, and whether the uncontrolled gas service data is uploaded accurately is judged. Specifically, assuming that the user has replaced the internet of things list and still performs offline payment, it can be inferred that the user is an elderly user, and an age tag corresponding to the elderly user is generated. And according to the single payment amount, the income level of the user family can be estimated, and a corresponding income level label is generated.
And deducing whether the user is living in the room for a long time according to the daily communication times of the terminal in the controlled gas service data for the label of whether the user is at home for a long time or not, if the daily communication times in the preset time period are zero, indicating that the user is not using gas for a long time.
For the user occupation label, the deduction can be obtained based on the amount and the frequency of single payment in the controlled gas service data. For example, if the amount of single recharging is small and the recharging frequency is high, the recharging time is fixed, and it can be inferred that the user is a renter who just works, and a corresponding user occupation label is generated. If the amount of the single recharging is large, the recharging frequency is high, and it can be inferred that the user may be a vendor, such as a user who opens a private career, a user who operates a small restaurant, and the like, and a corresponding user occupation label can be generated. If the amount of the single recharging is large, the household conditions of the users are good in frequency, and high-end products can be recommended directionally.
The above exemplary embodiments describe a method for verifying the inference of a part of the tags in the user behavior feature tag, and the following describes a method for verifying the inference of a part of the type of tags in the air consumption device feature tag:
For the tag of whether the gas water heater is used, whether the user is a gas water heater user or not can be estimated through daily gas consumption and gas consumption time in the controlled gas service data, the estimation result is compared with the uncontrolled gas service data, and whether the uncontrolled gas service data is uploaded accurately is judged.
For the label of the user kitchen range state, the kitchen range state of the user can be estimated through daily gas consumption and gas consumption time in the controlled gas service data. If the gas is used only in cooking time, the gas consumption per hour can be estimated according to the household population of the user and the gas consumption data, the gas consumption of the surrounding neighbors is combined, the pressure interference factor of the area is eliminated, whether the user is a user using an old kitchen range can be primarily judged, the inferred result is compared with the uncontrolled gas service data, and whether the uncontrolled gas service data is uploaded accurately is judged.
For the tag of whether the 5G is popular, the popularity of the 5G base station can be judged according to the average value of the signal intensity of the cell table in the controlled gas service data. Comparing the inferred result with the 5G base station popularity information provided by a third party data provider in the uncontrolled gas service data, and further judging whether the uncontrolled gas service data is uploaded accurately or not, and whether the data source is reliable or not. According to the popularity of the 5G base station, intelligent home products can be pertinently recommended to users in the later period, and the mining of potential customers and the popularization of the products are facilitated.
Further, a tag can be generated based on the uncontrolled gas service data, the tag is verified based on the controlled gas service data, and the credibility of the data source of the uncontrolled gas service data is judged based on the verification result. For example, based on uncontrolled gas service data obtained from a certain data source, a certain number of tags are generated from the uncontrolled gas service data, at least some of the tags are verified based on the controlled gas service data, and if the verification passing proportion is lower than a preset proportion, the uncontrolled gas service data obtained based on the data source is considered to be unreliable.
As an optional implementation manner, the user portrait construction method provided by the embodiment of the application further comprises the following steps:
Based on the user labels, the user groups are clustered through an improved K-means clustering algorithm, different user clusters are constructed, comprehensive images of the users are formed, and visual presentation of the results is completed.
Based on basic data such as user behavior feature tags, gas consumption characteristic feature tags, gas consumption equipment feature tags and the like, a data set is generated, and an improved K-means clustering algorithm is utilized to cluster user groups. The improved K-means clustering algorithm comprises the following calculation steps:
Step S201, randomly selecting a sample from the data set as an initial clustering center c 1;
Step S202, firstly, calculating the shortest distance between each sample and the current existing clustering center, wherein in the embodiment of the application, the shortest distance is represented by D (x);
the probability that each sample is selected as the next cluster center is then calculated as follows:
,
where D (x) represents the distance of the sample from the cluster center (referring to the selected cluster center), Representing the sum of Euclidean distances of all sample points from cluster center points of clusters to which the sample points belong;
Finally, selecting the next clustering center according to a wheel disc method;
Step S203, repeating the step S202 until K total cluster centers are selected;
Step S204, calculating the distance from each sample X i in the dataset to K clustering centers and dividing the distances into clusters corresponding to the clustering centers with the smallest distances;
Step S205, for each cluster c i, recalculate its cluster center, where the cluster center can be expressed by the following formula:
,
Where c i denotes the center of the ith cluster, |c i | is the number of data points in the ith cluster, Representing the sum of the values of all the data points in the ith cluster.
Step S206, repeating step S204 and step S205 until the position of the cluster center is not changed.
As shown in fig. 2 and 3, a schematic diagram of a user portrait provided by an embodiment of the present application is exemplarily shown. It can be seen that for user a, the revenue level is high and the risk factor is small, so that the method is suitable for being used as a directional recommendation user of high-end products. For the user B, the probability is that the elderly are living alone, the equipment is old, the risk coefficient of the user is high, and the security check service needs to be provided timely and regularly.
As an optional implementation manner, in step S14, automatically updating the user portrait according to the change of the gas service data, including periodically collecting the uncontrolled gas service data to update the user portrait, and/or receiving the uncontrolled gas service data actively uploaded by the gas user through a user uploading interface to update the user portrait. By the method, the user portrait is updated in time, the user image is enabled to be representative to the user, and further accuracy of a user analysis result generated based on the user portrait can be guaranteed.
For example, the gas consumer may have different types, numbers, and status of gas consumers at the time of gas consumer opening, and thus, previous consumer portraits may be difficult to represent today's consumers. In addition, it should be noted that, the data information such as the type, the number, the state and the like of the gas consumption equipment owned by the gas user belongs to the uncontrolled gas service data, and at present, the value of the data information is not fully utilized, and the influence of the uncontrolled gas service data on the user is not considered in the related technology, but only analyzed according to the controlled gas service data provided by the controlled internal system. According to the user portrait construction method provided by the embodiment of the application, the controlled gas service data and the uncontrolled gas service data are combined to generate the user portrait, so that the user portrait can be perfected, and the user portrait is more representative to gas users.
As shown in fig. 4, the embodiment of the application further provides an adaptive setting method for an alarm threshold, which includes the following steps:
s31, obtaining user portraits of a plurality of gas users, wherein the user portraits are generated by adopting the user portraits construction method provided by the embodiment of the application;
step S32, based on the user portrait, clustering the gas users to obtain different user clusters;
and step S33, setting an alarm threshold according to each cluster of gas users.
According to the above description, the method for adaptively setting the alarm threshold according to the embodiment of the application classifies the users according to the user images, and sets the alarm threshold for each class of users respectively, thereby reducing the probability of false alarm. The user portrait is generated by adopting the method for constructing the user portrait provided by the embodiment of the application, fully uses the values of the controlled gas service data and the uncontrolled gas service data, and is more representative to gas users.
As an optional implementation manner, the alarm threshold adaptive setting method provided by the embodiment of the application further includes the following steps:
based on the user portrait of each user cluster, acquiring a user behavior characteristic label and an air consumption equipment characteristic label of each user cluster;
based on the user behavior feature tag and the gas utilization equipment feature tag, extracting the basic attribute and the gas utilization equipment attribute of the gas user;
And setting an initial value of an alarm threshold value based on the basic attribute and the gas using equipment attribute and combining the maximum load flow of the gas using equipment.
Specifically, a user behavior feature tag and a gas consumption device feature tag of each cluster of users are obtained through user portraits, wherein the user behavior feature tag can comprise one or more of the following types of tags: user age group, home status, user occupation, user stove status, etc. The air device feature tag may comprise one or more of the following types of tags: whether a gas water heater is used, whether a gas warmer is used, whether 5g is popular, and whether security check requirements exist.
Based on the user behavior feature tag and the gas consumption equipment feature tag, the basic attribute and the gas consumption equipment attribute of the gas user can be extracted. For example, assume that the basic attributes of a clustered user include: the home state is daily regular work, and the user occupation is individual operators (such as operating restaurants, riding houses and the like); and, the gas consumption equipment attribute of the cluster user comprises: the gas water heater and the gas warmer are used. Then, for the clustered users, the initial value of the alarm threshold can be correspondingly increased so as to reduce the possibility of false alarm. Or assuming that the basic attributes of a clustered user include: old people, long-term families, old and old kitchen ranges and the like, gas equipment attributes comprise using a gas water heater, not using a gas warmer and the like, and for the clustered users, the initial value of the alarm threshold can be correspondingly reduced so as to discover possible abnormal gas consumption conditions earlier.
As an optional implementation manner, the alarm threshold adaptive setting method provided by the embodiment of the application further includes adjusting an alarm threshold update period. The method specifically comprises the following steps:
acquiring historical gas consumption in gas business data, sequencing the historical gas consumption according to a time sequence to obtain a gas consumption time sequence, and performing fast Fourier transform on the gas consumption time sequence;
calculating the amplitude of each frequency component, and finding out the frequency with the maximum amplitude after obtaining the amplitude of each frequency;
and converting the maximum frequency into a period length to obtain an estimated value of the threshold updating period.
Specifically, as shown in fig. 5, a gas consumption time series schematic diagram is exemplarily shown. The time series of gases were subjected to a fast fourier transform to obtain a spectral diagram as shown in fig. 6. And calculating the reciprocal of the maximum frequency, namely converting the maximum frequency into the period length, and obtaining the estimated value of the current threshold updating period. By means of the method, the updating period of the gas alarm threshold is dynamically adjusted, and the alarm accuracy and timeliness can be improved.
As an optional implementation manner, the alarm threshold adaptive setting method provided by the embodiment of the application further includes:
Establishing an LSTM neural network model, and training the LSTM neural network model based on the characteristic labels of the user portraits;
Predicting the gas consumption trend based on the trained LSTM neural network model;
based on the gas trend and the preset safety standard, judging whether the gas user has potential safety risk,
If the potential safety risk is judged, the alarm threshold value is reduced; and if no safety risk is judged, the alarm threshold value is improved. In this way, the sensitivity of the monitoring system to abnormal conditions is improved or terminal false alarms are reduced.
Specifically, a multidimensional differentiated user portrait analysis model construction technology is applied to extract and construct user portrait data features which are conducive to predicting gas consumption trend on a large scale, an LSTM neural network model is trained, and the gas consumption trend is predicted based on the trained LSTM neural network model. And (3) combining the predicted gas consumption trend with a preset safety standard, autonomously diagnosing whether the current alarm threshold is suitable or not, and marking the safety risk. For example, if a sudden large increase in the gas usage by a user is detected, and the change does not correspond to the user's normal gas usage habits, the monitoring system may mark this behavior as a potential safety risk.
And establishing a compensation mechanism based on self-learning, dynamically adjusting gas consumption prediction, and automatically adjusting an alarm threshold according to the safety risk. For example, if the gas consumption exceeds the normal range, the alarm threshold is lowered, so that the sensitivity of the system to abnormal conditions can be improved; if the gas consumption behavior of the user is kept stable for a long time, the alarm threshold can be improved to reduce false alarm of the terminal.
The above disclosure is illustrative of the preferred embodiments of the present application, but it should not be construed as limiting the scope of the application as will be understood by those skilled in the art: changes, modifications, substitutions, combinations, and simplifications may be made without departing from the spirit and scope of the application and the appended claims, and equivalents may be substituted and still fall within the scope of the application.
The technical application field of the invention includes but is not limited to gas safety, and the scope of the technical scheme of the embodiments of the invention is within the protection scope of the invention as long as the essence of the technical scheme does not deviate from the scope of the technical scheme of the embodiments of the invention.
Claims (10)
1. The user portrait construction method is characterized by comprising the following steps:
acquiring gas service data, wherein the gas service data comprises controlled gas service data and uncontrolled gas service data, and the controlled gas service data is derived from an internal controlled system data source of a service company; the uncontrolled gas service data is derived from a data source other than the internal controlled system data source;
identifying the gas service data through a pre-established label model to generate a user label;
summarizing the user labels to obtain user portraits;
And automatically updating the user portrait according to the change of the gas service data.
2. The method for constructing a user figure according to claim 1, wherein,
The tag model builds features based on the gas business data, the features being used to symbolize gas users, the tag model identifying the features to generate the user tags;
The user tag includes:
The gas consumption device comprises a user behavior characteristic tag, a gas consumption characteristic tag and a gas consumption equipment characteristic tag, wherein the user behavior characteristic tag is used for reflecting basic attributes of a gas user; the gas consumption characteristic feature tag is used for reflecting the gas consumption attribute of the gas user, and the gas consumption equipment characteristic tag is used for reflecting the gas consumption equipment attribute of the gas user;
The user behavior characteristic label, the gas consumption characteristic label and the gas consumption equipment characteristic label are directly obtained through the gas service data or are obtained through inference of the gas service data.
3. The method for constructing a user figure according to claim 2, wherein,
The basic attributes of the gas user include: lifestyle attributes, family structure attributes, economic capability attributes, and operational behavior attributes;
the gas utilization attributes of the gas users comprise: gas utilization behavior attributes and production operation attributes;
the gas utilization equipment attributes include: a gas usage type attribute and a gas usage parameter attribute.
4. The user portrait construction method according to claim 1, further comprising:
Based on the user labels, the user groups are clustered through an improved K-means clustering algorithm, different user clusters are constructed, and comprehensive user portraits are formed and visually presented.
5. The method for constructing a user figure according to claim 1, wherein,
The internally controlled system data source includes: an indoor gas utilization terminal, a work order security inspection system, a customer service and income system, a call telephone traffic system and a mobile client;
The uncontrolled gas business data sources include: social media, published network resources, third party data providers, and an internal unmanaged and audited database of the business company.
6. The method of constructing a user representation according to claim 1, further comprising:
And periodically collecting the uncontrolled gas service data to update the user portrait, and/or receiving the uncontrolled gas service data actively uploaded by a gas user through a user uploading interface to update the user portrait.
7. An alarm threshold adaptive setting method, which is characterized by comprising the following steps:
acquiring user portraits of a plurality of gas users, wherein the user portraits are generated based on the user portrayal construction method according to any one of claims 1 to 6;
Based on the user portrait, clustering the gas users to obtain different user clusters;
and setting an alarm threshold according to each cluster of gas users.
8. The method for adaptively setting an alarm threshold as in claim 7, wherein,
The user portrait includes a user behavior feature tag, a gas consumption feature tag and a gas consumption device feature tag,
The user behavior feature tag is used for reflecting basic attributes of the gas user; the gas consumption characteristic feature tag is used for reflecting the gas consumption attribute of the gas user, and the gas consumption equipment characteristic tag is used for reflecting the gas consumption equipment attribute of the gas user;
the method may further comprise the steps of,
Based on the user portraits of each user cluster, acquiring a user behavior characteristic label and an air utilization device characteristic label of each user cluster;
extracting the basic attribute and the gas utilization equipment attribute of a gas user based on the user behavior characteristic tag and the gas utilization equipment characteristic tag;
And setting an initial value of the alarm threshold value based on the basic attribute and the gas using equipment attribute and combining the maximum load flow of the gas using equipment.
9. The method for adaptively setting an alarm threshold as in claim 8, wherein said method comprises,
Acquiring historical gas consumption in gas business data, sequencing the historical gas consumption according to a time sequence to obtain a gas consumption time sequence, and performing fast Fourier transform on the gas consumption time sequence;
calculating the amplitude of each frequency component, and finding out the frequency with the maximum amplitude after obtaining the amplitude of each frequency;
and converting the maximum frequency into a period length to obtain an estimated value of the threshold updating period.
10. The method for adaptively setting an alarm threshold as in claim 9, further comprising,
Establishing an LSTM neural network model, and training the LSTM neural network model based on the characteristic labels of the user portraits;
predicting the gas consumption trend based on the trained LSTM neural network model;
based on the gas consumption trend and a preset safety standard, judging whether the gas user has potential safety risk,
If the potential safety risk is judged, the alarm threshold value is reduced; and if no safety risk is judged, the alarm threshold value is improved.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410411821.8A CN118014664B (en) | 2024-04-08 | User portrait construction method and alarm threshold self-adaptive setting method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410411821.8A CN118014664B (en) | 2024-04-08 | User portrait construction method and alarm threshold self-adaptive setting method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN118014664A true CN118014664A (en) | 2024-05-10 |
CN118014664B CN118014664B (en) | 2024-08-02 |
Family
ID=
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108596679A (en) * | 2018-04-27 | 2018-09-28 | 中国联合网络通信集团有限公司 | Construction method, device, terminal and the computer readable storage medium of user's portrait |
CN116501957A (en) * | 2023-03-28 | 2023-07-28 | 深圳兔展智能科技有限公司 | User tag portrait processing method, user portrait system, apparatus and storage medium |
CN117093883A (en) * | 2023-09-15 | 2023-11-21 | 成都秦川物联网科技股份有限公司 | Intelligent gas abnormal data analysis method, internet of things system, device and medium |
CN117350680A (en) * | 2023-11-28 | 2024-01-05 | 成都秦川物联网科技股份有限公司 | Intelligent gas safety management method based on user activity and Internet of things system |
CN117455529A (en) * | 2023-10-26 | 2024-01-26 | 中国电力科学研究院有限公司 | User electricity utilization characteristic image construction method and system based on big data technology |
CN117474330A (en) * | 2023-10-31 | 2024-01-30 | 中国联合网络通信集团有限公司 | Multi-disaster differentiation early warning information generation method and equipment considering user portraits |
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108596679A (en) * | 2018-04-27 | 2018-09-28 | 中国联合网络通信集团有限公司 | Construction method, device, terminal and the computer readable storage medium of user's portrait |
CN116501957A (en) * | 2023-03-28 | 2023-07-28 | 深圳兔展智能科技有限公司 | User tag portrait processing method, user portrait system, apparatus and storage medium |
CN117093883A (en) * | 2023-09-15 | 2023-11-21 | 成都秦川物联网科技股份有限公司 | Intelligent gas abnormal data analysis method, internet of things system, device and medium |
CN117455529A (en) * | 2023-10-26 | 2024-01-26 | 中国电力科学研究院有限公司 | User electricity utilization characteristic image construction method and system based on big data technology |
CN117474330A (en) * | 2023-10-31 | 2024-01-30 | 中国联合网络通信集团有限公司 | Multi-disaster differentiation early warning information generation method and equipment considering user portraits |
CN117350680A (en) * | 2023-11-28 | 2024-01-05 | 成都秦川物联网科技股份有限公司 | Intelligent gas safety management method based on user activity and Internet of things system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10896203B2 (en) | Digital analytics system | |
US10082779B2 (en) | Energy management computer system | |
US20180033087A1 (en) | Connected devices for detecting and responding to events in an environment | |
US20180010818A1 (en) | Forecast-based automatic scheduling of a distributed network of thermostats with learned adjustment | |
CN106575422B (en) | Energy management system and method | |
US11656097B2 (en) | Methods, systems, apparatuses and devices for optimizing utility consumption associated with at least one premises | |
EP3469496B1 (en) | Situation forecast mechanisms for internet of things integration platform | |
US20240209992A1 (en) | Method for determining a monitoring scheme and safety monitoring internet of things systems of pipe network reliability degree based on smart gas | |
CN109523276A (en) | The pre-warning monitoring method and device of customer complaint | |
Zhang et al. | Socioscope: Human relationship and behavior analysis in social networks | |
CN109634942A (en) | A kind of energy data exception judgment method and device | |
US12025327B2 (en) | Energy efficiency and comfort optimization through extreme weather adaptivity and AI | |
CN108475257B (en) | Processing remote meter read data to analyze consumption patterns | |
CN108304990B (en) | Power failure sensitivity pre-judging method and system | |
CN116823496A (en) | Intelligent insurance risk assessment and pricing system based on artificial intelligence | |
CN111917574A (en) | Social network topology model and construction method thereof, user confidence degree and intimacy degree calculation method and telecommunication fraud intelligent interception system | |
US20210350471A1 (en) | Systems and methods for utilizing data from electricity monitoring devices for analytics modeling | |
CN118014664B (en) | User portrait construction method and alarm threshold self-adaptive setting method | |
CN108234435A (en) | A kind of automatic testing method based on IP classification | |
CN118014664A (en) | User portrait construction method and alarm threshold self-adaptive setting method | |
US20220303759A1 (en) | Methods and systems for characterizing wireless devices | |
CN115982646A (en) | Multi-source test data management method and system based on cloud platform | |
CN113220545B (en) | Method and device for distributing work orders and electronic equipment | |
CN116226293A (en) | Method and system for generating and managing power customer portrait | |
Swarnalaxmi et al. | User activity analysis driven anomaly detection in cellular network |
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 |