CN110570032B - Initial weight optimization method and system for user behavior model - Google Patents
Initial weight optimization method and system for user behavior model Download PDFInfo
- Publication number
- CN110570032B CN110570032B CN201910785845.9A CN201910785845A CN110570032B CN 110570032 B CN110570032 B CN 110570032B CN 201910785845 A CN201910785845 A CN 201910785845A CN 110570032 B CN110570032 B CN 110570032B
- Authority
- CN
- China
- Prior art keywords
- user behavior
- success rate
- matching degree
- behavior model
- initial weight
- 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.)
- Expired - Fee Related
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 25
- 238000000034 method Methods 0.000 title claims abstract description 21
- 239000013598 vector Substances 0.000 claims abstract description 35
- 238000010276 construction Methods 0.000 claims description 5
- 238000007405 data analysis Methods 0.000 abstract 1
- 230000006399 behavior Effects 0.000 description 81
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
- H04L12/2803—Home automation networks
- H04L12/2823—Reporting information sensed by appliance or service execution status of appliance services in a home automation network
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Automation & Control Theory (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Development Economics (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an initial weight optimization method of a user behavior model, which relates to the technical field of data analysis and comprises the steps of setting corresponding initial weights for all characteristic items in a user behavior vector and constructing the user behavior model; acquiring a real success rate of each intelligent household device and generating a real success rate distribution; calculating according to the user behavior model to obtain a prediction success rate of the intelligent household equipment to be controlled and generating prediction success rate distribution; calculating the matching degree between the predicted success rate distribution and the real success rate distribution by adopting a minimized objective function; adjusting the initial weight according to the matching degree and a preset adjustment step length so as to optimize the user behavior model; and substituting the weight corresponding to the matching degree as the optimal weight into the user behavior model, and predicting the success rate of the intelligent household equipment to be controlled according to the user behavior model. According to the method and the device, the prediction success rate distribution of the user behavior model is fit with the real success rate distribution, and the prediction accuracy is effectively improved.
Description
Technical Field
The invention relates to the technical field of smart home, in particular to an initial weight optimization method and system of a user behavior model.
Background
Along with the improvement of living standard of people, the requirements of people on living environment are higher and higher, and how to maximally meet the comfort of people and the intelligent requirements of home furnishing is the target of intelligent home furnishing research. Most of the existing intelligent home systems are in a single mechanical automatic mode, and the defects of the existing intelligent home systems are mainly that the intelligent degree is low, and the user behaviors need to be analyzed and predicted according to the analysis result to achieve the purposes of intelligence and simplification of the user behaviors in order to achieve the real intelligence. Accurate prediction of user behavior is the key to realizing the intellectualization of the family living environment.
In the prior art, user behavior can be predicted by a visual recognition system based on video and voice, but a large amount of data accumulation is needed; or the user behavior is predicted through wearable equipment, but the wearable equipment needs to be worn by the user all the time, so that the cost is high; the user behavior can also be predicted by constructing a user behavior model, and the initial weight of the user behavior model is set by experience when being determined, so that the initial weight is blind and subjective and is often in error with the actual situation, and therefore, the result returned by the user behavior model may be greatly different from the true value, and the initial weight needs to be optimized so that the predicted result is fit with the true result.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an initial weight optimization method of a user behavior model, which comprises the steps of presetting a plurality of intelligent household devices, obtaining historical behavior data of each intelligent household device controlled by a user, and constructing a user behavior vector aiming at each control behavior in the historical behavior data, wherein the user behavior vector comprises a plurality of characteristic items;
the initial weight optimization method specifically includes:
step S1, setting corresponding initial weight for each feature item in the user behavior vector, and constructing a user behavior model according to each user behavior vector;
step S2, aiming at each intelligent household device, obtaining the real success rate of the intelligent household device controlled corresponding to each user behavior vector, and generating real success rate distribution according to each real success rate;
step S3, calculating according to the user behavior model to obtain the controlled prediction success rate of the intelligent household equipment corresponding to each user behavior vector, and generating prediction success rate distribution according to each prediction success rate;
step S4, calculating the matching degree between the predicted success rate distribution and the real success rate distribution by adopting a minimized objective function, and adding the matching degree into a matching degree data set;
Step S5, counting the number of elements of the matching degree in the matching degree data set, and comparing the number of elements with a preset element threshold:
if the number of the elements is smaller than the element threshold, the process goes to step S6;
if the number of the elements is not less than the element threshold, judging whether the matching degrees of the continuous preset times tend to be stable:
if yes, go to step S7;
if not, go to step S6;
the preset number of times is not greater than the element threshold;
step S6, adjusting the initial weight according to the matching degree and a preset adjustment step length, substituting the adjusted weight into the user behavior model corresponding to the intelligent home equipment to optimize the user behavior model, and then turning to the step S3;
and S7, substituting the weight corresponding to the matching degree as an optimal weight into the user behavior model corresponding to the intelligent home equipment, and predicting the success rate of the intelligent home equipment being controlled according to the user behavior model.
Preferably, the feature items include occurrence time of the control action, operation state data of each smart home device, indoor environment data, and outdoor environment data.
Preferably, in step S2, the true success rate is 0 or 1.
Preferably, in step S3, the value range of the prediction success rate is [0,1 ].
Preferably, the prediction success rate is calculated according to the following formula:
wherein,
s (b) for indicating said prediction success rate;
c is used for representing the characteristic item;
w(c)for representing the weights.
Preferably, in step S4, the minimization objective function is KL divergence.
Preferably, in step S6, the preset adjustment step is 0.01.
An initial weight optimization system of a user behavior model, which applies any one of the above initial weight optimization methods of the user behavior model, the initial weight optimization system specifically includes:
the model building module is used for setting corresponding initial weights for the feature items in the user behavior vectors and building a user behavior model according to the user behavior vectors;
the data acquisition module is connected with the model construction module and is used for acquiring the real success rate corresponding to the intelligent home equipment corresponding to each user behavior vector aiming at each intelligent home equipment and generating real success rate distribution according to each real success rate;
The first processing module is connected with the model construction module and used for calculating to obtain the controlled prediction success rate of the intelligent household equipment corresponding to each user behavior vector according to the user behavior model and generating prediction success rate distribution according to each prediction success rate;
the second processing module is respectively connected with the data acquisition module and the first processing module and is used for calculating the matching degree between the predicted success rate distribution and the real success rate distribution by adopting a minimized objective function and adding the matching degree into a matching degree data set;
the first judgment module is connected with the second processing module and used for counting the number of elements of the matching degree in the matching degree data set, generating and outputting a corresponding first judgment result when the number of the elements is smaller than a preset element threshold value, and generating and outputting a corresponding second judgment result when the number of the elements is not smaller than the element threshold value;
the second judgment module is connected with the first judgment module and used for judging whether each matching degree of continuous preset times tends to be stable according to the second judgment result, outputting a corresponding third judgment result when each matching degree tends to be stable, and outputting a corresponding fourth judgment result when each matching degree does not tend to be stable;
The data adjusting module is respectively connected with the first judging module and the second judging module, and specifically comprises:
the first data adjusting unit is used for adjusting the initial weight according to the first judgment result and the matching degree and a preset adjusting step length, and substituting the adjusted weight into the user behavior model to optimize the user behavior model;
the second data adjusting unit is used for adjusting the initial weight according to the fourth judgment result and the matching degree and a preset adjusting step length, and substituting the adjusted weight into the user behavior model corresponding to the intelligent home equipment so as to optimize the user behavior model;
and the model generation module is connected with the second judgment module and used for substituting the weight corresponding to the matching degree as an optimal weight into the user behavior model corresponding to the intelligent household equipment according to the third judgment result and predicting the success rate of controlling the intelligent household equipment according to the user behavior model.
The technical scheme has the following advantages or beneficial effects: by optimizing the initial weight, the prediction success rate distribution of the user behavior model is fit with the real success rate distribution, and the prediction accuracy of the user behavior model is effectively improved.
Drawings
FIG. 1 is a flow chart illustrating a method for optimizing initial weights of a user behavior model according to a preferred embodiment of the present invention;
FIG. 2 is a schematic structural diagram of an initial weight optimization system for a user behavior model according to a preferred embodiment of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present invention is not limited to the embodiment, and other embodiments may be included in the scope of the present invention as long as the gist of the present invention is satisfied.
In a preferred embodiment of the present invention, based on the above problems in the prior art, a method for optimizing initial weights of a user behavior model is provided, where a plurality of smart home devices are preset, historical behavior data of each smart home device controlled by a user is obtained, a user behavior vector is constructed for each control behavior in the historical behavior data, and the user behavior vector includes a plurality of feature items;
as shown in fig. 1, the initial weight optimization method specifically includes:
step S1, setting corresponding initial weight for each feature item in the user behavior vector, and constructing a user behavior model according to each user behavior vector;
Step S2, aiming at each intelligent household device, obtaining the real success rate of the intelligent household device controlled corresponding to each user behavior vector, and generating real success rate distribution according to the real success rate;
step S3, calculating according to the user behavior model to obtain the controlled prediction success rate of the intelligent household equipment corresponding to each user behavior vector, and generating prediction success rate distribution according to each prediction success rate;
step S4, calculating the matching degree between the predicted success rate distribution and the real success rate distribution by adopting a minimized objective function, and adding the matching degree into a matching degree data set;
step S5, counting the number of elements of the matching degree in the matching degree data set, and comparing the number of elements with a preset element threshold:
if the number of elements is less than the element threshold, go to step S6;
if the number of the elements is not less than the element threshold, judging whether the matching degrees of the continuous preset times tend to be stable:
if yes, go to step S7;
if not, go to step S6;
the preset times are not greater than an element threshold;
step S6, adjusting the initial weight according to the matching degree and a preset adjustment step length, substituting the adjusted weight into a user behavior model corresponding to the intelligent household equipment to optimize the user behavior model, and then turning to step S3;
And step S7, substituting the weight corresponding to the matching degree as the optimal weight into the user behavior model corresponding to the intelligent household equipment, and predicting the success rate of controlling each intelligent household equipment according to the user behavior model.
Specifically, in the embodiment, by optimizing the initial weight, the difference between the actual success rate distribution of the smart home devices being controlled and the predicted success rate distribution obtained by calculating the user behavior model before the actual success rate distribution is effectively shortened, and the prediction accuracy of the user behavior model is effectively improved. Specifically, in the historical behavior data, if the smart home devices are controlled, the corresponding real success rate is 1, and if the smart home devices are not controlled, the corresponding real success rate is 0. The prediction success rate distribution is infinitely close to the real success rate distribution, but is not equal to the real success rate distribution, so the value range of the prediction success rate is between 0 and 1, when the possibility that the intelligent household equipment is controlled is high, the prediction success rate is infinitely close to 1, and when the possibility that the intelligent household equipment is controlled is low, the prediction success rate is infinitely close to 0.
In this embodiment, each smart home device corresponds to a corresponding user behavior model, for each smart home device, a prediction success rate of the smart home device is first obtained through a preset initial weight, a prediction success rate distribution is generated according to the prediction success rate, and a matching degree between the prediction success rate distribution and a true success rate distribution is calculated. After multiple times of iterative optimization, the prediction success rate of the intelligent home equipment is obtained according to the optimized weight processing, the prediction success rate distribution is generated according to the prediction success rate, the matching degree between the prediction success rate distribution and the real success rate distribution is calculated, and the matching degrees obtained continuously and repeatedly are compared, so that whether the matching degree between the prediction success rate distribution and the real success rate distribution tends to be stable or not can be seen, if the matching degree tends to be stable, the weight corresponding to the finally obtained matching degree is used as the optimal weight, if the matching degree does not tend to be stable, the iterative optimization is continuously carried out according to the method, and preferably, the iterative optimization times are not less than ten thousand times.
In a preferred embodiment of the present invention, the characteristic items include occurrence time of a control action, operation state data of each smart home device, indoor environment data, and outdoor environment data.
In the preferred embodiment of the present invention, in step S2, the true success rate is 0 or 1.
In the preferred embodiment of the present invention, in step S3, the value range of the prediction success rate is [0,1 ].
In a preferred embodiment of the present invention, the prediction success rate is calculated according to the following formula:
wherein,
s (b) for indicating a prediction success rate;
c is used for representing the characteristic items;
w(c)for representing the weights.
In the preferred embodiment of the present invention, in step S4, the minimized objective function is KL divergence.
In the preferred embodiment of the present invention, in step S6, the preset adjustment step size is 0.01.
An initial weight optimization system of a user behavior model, which applies any one of the above initial weight optimization methods of a user behavior model, as shown in fig. 2, specifically includes:
the model building module 1 is used for setting corresponding initial weights for all feature items in the user behavior vectors and building a user behavior model according to all the user behavior vectors;
The data acquisition module 2 is connected with the model construction module 1 and is used for acquiring the real success rate of the intelligent home equipment controlled corresponding to each user behavior vector aiming at each intelligent home equipment and generating real success rate distribution according to each real success rate;
the first processing module 3 is connected with the model building module 1 and used for calculating the controlled prediction success rate of the intelligent home equipment corresponding to each user behavior vector according to the user behavior model and generating prediction success rate distribution according to each prediction success rate;
the second processing module 4 is respectively connected with the data acquisition module 2 and the first processing module 3 and is used for calculating the matching degree between the predicted success rate distribution and the actual success rate distribution by adopting a minimized objective function and adding the matching degree into a matching degree data set;
the first judgment module 5 is connected with the second processing module 4 and used for counting the number of elements of the matching degree in the matching degree data set, generating and outputting a corresponding first judgment result when the number of the elements is smaller than a preset element threshold value, and generating and outputting a corresponding second judgment result when the number of the elements is not smaller than the element threshold value;
the second judgment module 6 is connected with the first judgment module 5 and used for judging whether each matching degree of the continuous preset times tends to be stable according to the second judgment result, outputting a corresponding third judgment result when each matching degree tends to be stable, and outputting a corresponding fourth judgment result when each matching degree does not tend to be stable;
The data adjusting module 7 is respectively connected to the first judging module 5 and the second judging module 6, and the data adjusting module 7 specifically includes:
the first data adjusting unit 71 is configured to adjust the initial weight according to the first determination result and the matching degree and according to a preset adjustment step length, and substitute the adjusted weight into a user behavior model corresponding to the smart home device to optimize the user behavior model;
the second data adjusting unit 72 is configured to adjust the initial weight according to the fourth determination result and the matching degree and according to a preset adjustment step length, and substitute the adjusted weight into a user behavior model corresponding to the smart home device to optimize the user behavior model;
and the model generation module 8 is connected with the second judgment module 6 and is used for substituting the weight corresponding to the matching degree as the optimal weight into the user behavior model corresponding to the intelligent home equipment according to the third judgment result and predicting the success rate of controlling each intelligent home equipment according to the user behavior model.
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.
Claims (6)
1. The initial weight optimization method of the user behavior model is characterized in that a plurality of intelligent household devices are preset, historical behavior data of the intelligent household devices controlled by a user are obtained, a user behavior vector is constructed for each control behavior in the historical behavior data, and the user behavior vector comprises at least one characteristic item;
the initial weight optimization method specifically includes:
step S1, setting corresponding initial weight for each feature item in the user behavior vector, and constructing a user behavior model according to each user behavior vector;
step S2, aiming at each intelligent household device, obtaining the real success rate of the intelligent household device controlled corresponding to each user behavior vector, and generating real success rate distribution according to each real success rate;
step S3, calculating according to the user behavior model to obtain the controlled prediction success rate of the intelligent household equipment corresponding to each user behavior vector, and generating prediction success rate distribution according to each prediction success rate;
step S4, calculating the matching degree between the predicted success rate distribution and the real success rate distribution by adopting a minimized objective function, and adding the matching degree into a matching degree data set;
Step S5, counting the number of elements of the matching degree in the matching degree data set, and comparing the number of elements with a preset element threshold:
if the number of the elements is smaller than the element threshold, turning to step S6;
if the number of the elements is not less than the element threshold, judging whether the matching degrees of the continuous preset times tend to be stable:
if yes, go to step S7;
if not, go to step S6;
the preset number of times is not greater than the element threshold;
step S6, adjusting the initial weight according to the matching degree and a preset adjustment step length, substituting the adjusted weight into the user behavior model corresponding to the intelligent household equipment to optimize the user behavior model, and turning to the step S3;
step S7, substituting the weight corresponding to the matching degree as an optimal weight into the user behavior model corresponding to the intelligent home equipment, and predicting the success rate of controlling the intelligent home equipment according to the user behavior model;
in the step S3, the value range of the prediction success rate is [0, 1 ];
calculating the prediction success rate according to the following formula:
Wherein,
s (b) is used to represent the prediction success rate;
c is used for representing the characteristic item;
w(c)for representing the weights.
2. The initial weight optimization method according to claim 1, wherein the characteristic items include occurrence time of the control action, operation state data of each smart home device, indoor environment data, and outdoor environment data.
3. The initial weight optimization method according to claim 1, wherein in the step S2, the true success rate is 0 or 1.
4. The initial weight optimization method according to claim 1, wherein in step S4, the minimization objective function is KL divergence.
5. The initial weight optimization method according to claim 1, wherein in the step S6, the preset adjustment step size is 0.01.
6. An initial weight optimization system of a user behavior model, characterized in that, the initial weight optimization method of the user behavior model according to any one of claims 1-5 is applied, and the initial weight optimization system specifically comprises:
the model building module is used for setting corresponding initial weights for the feature items in the user behavior vectors and building a user behavior model according to the user behavior vectors;
The data acquisition module is connected with the model construction module and is used for acquiring the real success rate of the intelligent home equipment controlled corresponding to each user behavior vector aiming at each intelligent home equipment and generating real success rate distribution according to each real success rate;
the first processing module is connected with the model construction module and used for calculating to obtain the controlled prediction success rate of the intelligent household equipment corresponding to each user behavior vector according to the user behavior model and generating prediction success rate distribution according to each prediction success rate;
the second processing module is respectively connected with the data acquisition module and the first processing module and is used for calculating the matching degree between the predicted success rate distribution and the real success rate distribution by adopting a minimized objective function and adding the matching degree into a matching degree data set;
the first judgment module is connected with the second processing module and used for counting the number of elements of the matching degree in the matching degree data set, generating and outputting a corresponding first judgment result when the number of the elements is smaller than a preset element threshold value, and generating and outputting a corresponding second judgment result when the number of the elements is not smaller than the element threshold value;
The second judgment module is connected with the first judgment module and used for judging whether each matching degree for continuous preset times tends to be stable or not according to the second judgment result, outputting a corresponding third judgment result when each matching degree tends to be stable, and outputting a corresponding fourth judgment result when each matching degree does not tend to be stable;
the data adjusting module is respectively connected with the first judging module and the second judging module, and specifically comprises:
the first data adjusting unit is used for adjusting the initial weight according to the first judgment result and the matching degree and a preset adjusting step length, and substituting the adjusted weight into the user behavior model corresponding to the intelligent home equipment so as to optimize the user behavior model;
the second data adjusting unit is used for adjusting the initial weight according to the fourth judgment result and the matching degree and a preset adjusting step length, and substituting the adjusted weight into the user behavior model corresponding to the intelligent home equipment so as to optimize the user behavior model;
the model generation module is connected with the second judgment module and used for substituting the weight corresponding to the matching degree as an optimal weight into the user behavior model corresponding to the intelligent home equipment according to the third judgment result and predicting the success rate of controlling the intelligent home equipment according to the user behavior model;
The value range of the prediction success rate is [0, 1 ];
calculating the prediction success rate according to the following formula:
wherein,
s (b) for indicating said prediction success rate;
c is used for representing the characteristic item;
w(c)for representing the weights.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910785845.9A CN110570032B (en) | 2019-08-23 | 2019-08-23 | Initial weight optimization method and system for user behavior model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910785845.9A CN110570032B (en) | 2019-08-23 | 2019-08-23 | Initial weight optimization method and system for user behavior model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110570032A CN110570032A (en) | 2019-12-13 |
| CN110570032B true CN110570032B (en) | 2022-06-10 |
Family
ID=68775934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910785845.9A Expired - Fee Related CN110570032B (en) | 2019-08-23 | 2019-08-23 | Initial weight optimization method and system for user behavior model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110570032B (en) |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015173388A2 (en) * | 2014-05-15 | 2015-11-19 | Essilor International (Compagnie Generale D'optique) | A monitoring system for monitoring head mounted device wearer |
| CN106022505A (en) * | 2016-04-28 | 2016-10-12 | 华为技术有限公司 | Method and device of predicting user off-grid |
| US10252145B2 (en) * | 2016-05-02 | 2019-04-09 | Bao Tran | Smart device |
| CN109818839B (en) * | 2019-02-03 | 2022-02-25 | 三星电子（中国）研发中心 | Personalized behavior prediction method, device and system applied to smart home |
-
2019
- 2019-08-23 CN CN201910785845.9A patent/CN110570032B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN110570032A (en) | 2019-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11144616B2 (en) | Training distributed machine learning with selective data transfers | |
| KR102267180B1 (en) | Air conditioning system and controlling method thereof | |
| US20160282822A1 (en) | Control system with response time estimation | |
| JP7279445B2 (en) | Prediction method, prediction program and information processing device | |
| US20200200414A1 (en) | Smart home air conditioner automatic control system based on artificial intelligence | |
| CN111609534B (en) | Temperature control method and device and central temperature control system | |
| CN110769453B (en) | Multi-modal monitoring data dynamic compression control method under unstable network environment | |
| CN114065863B (en) | Federal learning method, apparatus, system, electronic device and storage medium | |
| CN111860789A (en) | Model training method, terminal and storage medium | |
| WO2015095238A1 (en) | Occupancy detection | |
| CN111256307A (en) | Temperature control method, air conditioning apparatus, control apparatus, and storage medium | |
| CN114838470A (en) | Control method and system for heating, ventilating and air conditioning | |
| CN110570032B (en) | Initial weight optimization method and system for user behavior model | |
| CN111242266A (en) | Operation data management system | |
| CN116753562B (en) | Graphene electric heating intelligent temperature control system based on data analysis | |
| CN110769000A (en) | Dynamic compression prediction control method of continuous monitoring data in unstable network transmission | |
| CN112394647A (en) | Control method, device and equipment of household equipment and storage medium | |
| CN110289090A (en) | Event finds method and device, storage medium, terminal | |
| US10761200B1 (en) | Method for evaluating positioning parameters and system | |
| JP6559182B2 (en) | Control system for response time estimation and automatic operating parameter adjustment | |
| CN115218358B (en) | Indoor air environment adjusting method and equipment | |
| CN109035178A (en) | A kind of multi-parameter value tuning method applied to image denoising | |
| CN114576816A (en) | Air conditioner adjusting method and device based on infrared sensor | |
| CN114025321A (en) | Massive power internet of things terminal access control method based on rapid uplink authorization | |
| EP3570498B1 (en) | Adaptive hvac control system |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220610 |