CN110409955A - Window control method, device and electronic equipment based on decision tree prediction model - Google Patents
Window control method, device and electronic equipment based on decision tree prediction model Download PDFInfo
- Publication number
- CN110409955A CN110409955A CN201910492486.8A CN201910492486A CN110409955A CN 110409955 A CN110409955 A CN 110409955A CN 201910492486 A CN201910492486 A CN 201910492486A CN 110409955 A CN110409955 A CN 110409955A
- Authority
- CN
- China
- Prior art keywords
- data
- decision tree
- prediction model
- window
- tree prediction
- 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.)
- Pending
Links
- 238000003066 decision tree Methods 0.000 title claims abstract description 112
- 238000000034 method Methods 0.000 title claims abstract description 48
- 230000007613 environmental effect Effects 0.000 claims abstract description 24
- 238000004140 cleaning Methods 0.000 claims description 8
- 230000005855 radiation Effects 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 6
- 239000013589 supplement Substances 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000003860 storage Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 2
- 238000005406 washing Methods 0.000 claims 1
- 238000009423 ventilation Methods 0.000 abstract description 8
- 230000008569 process Effects 0.000 description 10
- 238000010586 diagram Methods 0.000 description 6
- 238000005070 sampling Methods 0.000 description 5
- 238000005265 energy consumption Methods 0.000 description 3
- 241000238876 Acari Species 0.000 description 2
- 241001269238 Data Species 0.000 description 2
- 238000004378 air conditioning Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000003750 conditioning effect Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000004992 fission Effects 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000032696 parturition Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- E—FIXED CONSTRUCTIONS
- E05—LOCKS; KEYS; WINDOW OR DOOR FITTINGS; SAFES
- E05F—DEVICES FOR MOVING WINGS INTO OPEN OR CLOSED POSITION; CHECKS FOR WINGS; WING FITTINGS NOT OTHERWISE PROVIDED FOR, CONCERNED WITH THE FUNCTIONING OF THE WING
- E05F15/00—Power-operated mechanisms for wings
- E05F15/70—Power-operated mechanisms for wings with automatic actuation
-
- E—FIXED CONSTRUCTIONS
- E05—LOCKS; KEYS; WINDOW OR DOOR FITTINGS; SAFES
- E05F—DEVICES FOR MOVING WINGS INTO OPEN OR CLOSED POSITION; CHECKS FOR WINGS; WING FITTINGS NOT OTHERWISE PROVIDED FOR, CONCERNED WITH THE FUNCTIONING OF THE WING
- E05F15/00—Power-operated mechanisms for wings
- E05F15/70—Power-operated mechanisms for wings with automatic actuation
- E05F15/71—Power-operated mechanisms for wings with automatic actuation responsive to temperature changes, rain, wind or noise
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
- G05B23/0245—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model based on a qualitative model, e.g. rule based; if-then decisions
- G05B23/0248—Causal models, e.g. fault tree; digraphs; qualitative physics
-
- E—FIXED CONSTRUCTIONS
- E05—LOCKS; KEYS; WINDOW OR DOOR FITTINGS; SAFES
- E05Y—INDEXING SCHEME ASSOCIATED WITH SUBCLASSES E05D AND E05F, RELATING TO CONSTRUCTION ELEMENTS, ELECTRIC CONTROL, POWER SUPPLY, POWER SIGNAL OR TRANSMISSION, USER INTERFACES, MOUNTING OR COUPLING, DETAILS, ACCESSORIES, AUXILIARY OPERATIONS NOT OTHERWISE PROVIDED FOR, APPLICATION THEREOF
- E05Y2900/00—Application of doors, windows, wings or fittings thereof
- E05Y2900/10—Application of doors, windows, wings or fittings thereof for buildings or parts thereof
- E05Y2900/13—Type of wing
- E05Y2900/148—Windows
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Air Conditioning Control Device (AREA)
Abstract
The present invention relates to window control method, device and electronic equipments based on decision tree prediction model, comprising: obtains indoor occupant to the control data and environmental data of window;Environmental data includes outdoor environment data;Control data and outdoor environment data are handled, to obtain treated data;Based on treated data, decision tree prediction model is generated;According to decision tree prediction model, the state of output control signal control window.The present invention, to the control data and environmental data of window, establishes decision tree prediction model by indoor occupant, and realizes the control to window state based on decision tree prediction model, is finally reached the purpose for improving indoor occupant thermal comfort using gravity-flow ventilation.
Description
Technical field
The present invention relates to indoor environmental condition control fields, more specifically to a kind of window based on decision tree prediction model
Family control method, device and electronic equipment.
Background technique
Compared to force ventilation, the characteristics of natural wind is due to its ripple frequency, it is easier to take away heat from human body, improve people
Body heat balance;Simultaneously because long-term evolution as a result, the mankind psychologically more have a preference for natural wind.In addition, using gravity-flow ventilation,
The use of air-conditioning can also be reduced, and then reduces energy consumption.It is naturally logical with the continuous development of Building Environment and Equipment Engineering industry
Air control system has become the important composition of the building management system (Building Management System, BMS) of modern architecture
Part.The purpose of gravity-flow ventilation control is gravity-flow ventilation to be controlled, to improve thermal comfort by close window.
However, most of window control systems are only unified management to window with electric window-opening equipment, not
Really realize the autocontrol switch control of window.Although part window control system is controlled using close window, control is usual
Only judged according to simple indoor/outdoor temperature-difference to realize, and cannot really reflect indoor occupant and heat is improved using gravity-flow ventilation and is relaxed
The wish of adaptive, and cannot be guaranteed to realize the effect for meeting indoor occupant thermal comfort.
Summary of the invention
The technical problem to be solved in the present invention is that in view of the above drawbacks of the prior art, providing a kind of based on decision tree
Window control method, device and the electronic equipment of prediction model.
The technical solution adopted by the present invention to solve the technical problems is: providing a kind of window based on decision tree prediction model
Family control method, comprising:
Indoor occupant is obtained to the control data and environmental data of window;The environmental data includes outdoor environment data;
The control data and the outdoor environment data are handled, to obtain treated data;
Based on treated the data, decision tree prediction model is generated;
According to the decision tree prediction model, the state of output control signal control window.
In one embodiment, the indoor occupant includes the control data bag of window: window turn-on data and closing number
According to;
The location information of the window is obtained by photoelectric sensor or position sensor detection;
The outdoor environment data include: outdoor temperature, outside relative humidity, outside atmospheric pressure, outdoor radiation intensity,
Outdoor wind speed, outdoor wind direction and outdoor rainfall.
In one embodiment, described that the control data and the outdoor environment data are handled, to be located
Data after reason include:
With the acquisition time of the control data and outdoor environment data sequence, to the control data and the room
External environment data are arranged;
By after arrangement control data and outdoor environment data be stored in database profession.
In one embodiment, after being stored in database profession the control data after arranging and outdoor environment data also
Include:
The data being stored in database profession are cleaned, with the data after being cleaned;
Data after cleaning are split, to obtain the segmentation data for generating the decision tree prediction model;Institute
Stating segmentation data is treated the data.
In one embodiment, the described pair of data being stored in database profession, which clean, includes:
The data of value missing are found out, and replace the value of missing, the data lacked with supplement value using interpolation method;
Using method of analysis of variance, value mistake or abnormal data are identified;
The data of described value mistake or exception are removed, and replace the value of missing using interpolation method, to supplement described value mistake
Or abnormal data;
Eliminate the data for repeating record.
In one embodiment, the data after described pair of cleaning are split, pre- for generating the decision tree to obtain
Survey model segmentation data include:
Data after cleaning are equally divided by first part's data and second part data using random fashion, wherein institute
First part's data are stated for being trained to the decision tree prediction model, the second part data are used for the decision
Tree prediction model is verified.
In one embodiment, described based on treated the data, generating decision tree prediction model includes:
First part's data are pre-processed, the input data and output data of decision tree prediction model are obtained;
The input data includes: outdoor temperature, outside relative humidity, outside atmospheric pressure, outdoor radiation intensity, outdoor wind speed, room
Outer wind direction and outdoor rainfall;The output data is the state of window;
The input data and output data are imported, generating includes multiple and different first decision tree prediction models for setting layers;
According to the tree layer depth of decision tree and classification accuracy relationship, the number of plies of the decision tree prediction model is determined;
Based on the identified number of plies, the decision tree prediction model is generated.
In one embodiment, described to import the input data and output data, generating includes multiple and different tree layers
Just decision tree prediction model includes:
The input data is handled based on C4.5 algorithm, obtains the information of each feature in the input data
The ratio of gains;
The information gain ratio of each feature is compared judgement, obtains the maximum first information ratio of gains;
According to the maximum first information ratio of gains, determining spy corresponding with the maximum first information ratio of gains
Sign;
Using feature corresponding with the maximum first information ratio of gains as the decision point of each node of the decision tree;
Calculate the information gain ratio of each characteristic value in the decision point;
The information gain ratio of each characteristic value is compared judgement, obtains maximum second information gain ratio;
According to the maximum second information gain ratio, determine with maximum second information gain than corresponding feature
Value;
Using with maximum second information gain than corresponding characteristic value further dividing as the decision point
Cut-point;
Input data is divided in the node based on the decision point and the cut-point;
It repeats the above steps, constantly divides each node, extend the tree layer depth of decision tree until presetting the number of plies, to generate
The just decision tree prediction model.
The present invention also provides a kind of window control devices based on decision tree prediction model, comprising:
Acquiring unit, for obtaining indoor occupant to the control data and environmental data of window;The environmental data includes
Outdoor environment data;
Processing unit, for handling the control data and the outdoor environment data, to obtain, treated
Data;
Execution unit, for generating decision tree prediction model based on treated the data;
Control unit, for according to the decision tree prediction model, the state of output control signal control window.
The present invention also provides a kind of electronic equipment, including memory and processor, the processor according to for depositing
The program instruction that reservoir is stored executes the step of approach described above.
The present invention also provides a kind of storage mediums, are stored thereon with computer program, the computer program is by processor
The step of method as described above is realized when execution.
Implement the window control method of the invention based on decision tree prediction model, has the advantages that include: to obtain
Take indoor occupant to the control data and environmental data of window;Environmental data includes outdoor environment data;To control data and room
External environment data are handled, to obtain treated data;Based on treated data, decision tree prediction model is generated;Root
According to decision tree prediction model, the state of output control signal control window.The present invention is by indoor occupant to the control number of window
According to and environmental data, establish decision tree prediction model, and the control to window state is realized based on decision tree prediction model, finally
Achieve the purpose that improve indoor occupant thermal comfort using gravity-flow ventilation.
Detailed description of the invention
Present invention will be further explained below with reference to the attached drawings and examples, in attached drawing:
Fig. 1 is the process signal of window control method one embodiment provided by the invention based on decision tree prediction model
Figure;
Fig. 2 is the process signal of window control method another embodiment provided by the invention based on decision tree prediction model
Figure;
Fig. 3 is the process signal of the window control method another embodiment provided by the invention based on decision tree prediction model
Figure;
Fig. 4 is the process signal of the window control method another embodiment provided by the invention based on decision tree prediction model
Figure;
Fig. 5 is the schematic diagram provided by the invention based on decision tree prediction model;
Fig. 6 is tree layer depth schematic diagram provided by the invention;
Fig. 7 is the structural schematic diagram of the window control device provided in an embodiment of the present invention based on decision tree prediction model;
Fig. 8 is the structural schematic diagram of electronic equipment provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Window control method provided by the present invention based on decision tree prediction model is by collecting indoor occupant to window
Control data and environmental data, after the two is associated, establish window state using C4.5 algorithm and information gain principle
The decision-tree model of control can promote gravity-flow ventilation using the optimization window control strategy of heuristic decision, improve Indoor Thermal
Comfort reduces energy consumption.
Specifically, being the window control method provided in an embodiment of the present invention based on decision tree prediction model with reference to Fig. 1
Flow diagram.
As shown in Figure 1, being somebody's turn to do the window control method based on decision tree prediction model includes: step S10, step S20, step
S30 and step S40.
Step S10, indoor occupant is obtained to the control data and environmental data of window.
Wherein, indoor occupant includes window turn-on data to the control data bag of window and closes data, i.e. indoor occupant pair
The control data of the switch state of window.Wherein, indoor occupant can be by being arranged at window to the control data of window
Photoelectric sensor or position sensor, which directly detect, to be collected.
Further, indoor occupant is adopted to the acquisition of the control data of window by the way of circle collection with default
Sample interval is acquired control data of the indoor occupant to window, and therefore, the indoor occupant of the embodiment of the present invention is to window
Control data are a data group, include each sampling instant control data collected.
In addition, if the sampling interval is Δ K, then indoor occupant is to window since the thermal conditioning period of human body is 15 minutes
Control data are referred to the thermal conditioning period of human body, and general setting is not more than 15 minutes, moreover, in order to reduce data acquisition
Amount, sampling interval should not also be arranged too small, generally can be set to 10 minutes or so.
Further, environmental data includes outdoor environment data.Wherein, outdoor environment data can use outdoor weather station
Data, wherein outdoor weather station data are equivalent to the data acquisition platform of outdoor various environmental parameters.The outdoor environment data packet
Include but be not limited to outdoor temperature, outdoor relative temperature, outside atmospheric pressure, outdoor radiation intensity, outdoor wind speed, outdoor wind direction and
Outdoor rainfall etc..Wherein, different environmental datas can be detected by corresponding sensor and be obtained.
It should be noted that usually indoor heat load generation be it is more stable, a steady state value is essentially, so can
With without the concern for indoor environment data, and the variation of indoor heat load caused by outdoor environment changes is bigger, therefore, the present invention
It is main to consider outdoor environment data.It will of course be understood that ground, in some other embodiment, it is also contemplated that indoor environment number
According to when increasing indoor environment data, principle is similar with the present invention.
It is synchronous with outdoor environment data similarly, for acquiring also by the way of circle collection for outdoor environment data
Acquisition, sampling interval are also configured as Δ K.
Step S20, control data and outdoor environment data are handled, to obtain treated data.
Specifically, as shown in Fig. 2, handling control data and outdoor environment data, in step S20 to be located
Data after reason specifically execute following processing movement:
Step S201, to control the acquisition time sequence of data and outdoor environment data, to control data and outdoor environment
Data are arranged.
Step S202, by after arrangement control data and outdoor environment data be stored in database profession.
Wherein, to control the acquisition time sequence of data and outdoor environment data, to control data and outdoor environment data
It is arranged are as follows: collected control data and outdoor environment data are organized into number according to respective acquisition time sequence respectively
It is stored in database profession according to group, and by the data group arranged.By taking the sampling interval is 10 minutes as an example, data group is stored in data
The storage form in library is as shown in the table:
Further, as shown in figure 3, further including step S203 and step S204 after step S202.
Step S203, the data being stored in database profession are cleaned, with the data after being cleaned.
Step S204, the data after cleaning are split, to obtain the segmentation number for generating decision tree prediction model
According to;Dividing data is treated data.
Specifically, being cleaned mainly the data being stored in database profession to deficiency of data, the mistake in database
Accidentally the data of value, the data for repeating to record are handled.
Wherein, for incomplete data, that is, it is worth the data of missing, the value of missing can be replaced using interpolation method, mends
Supplement the data of missing with money.For the data of error value, it is possible, firstly, to using method of analysis of variance, identify error value data or
The data of person's exceptional value, and then the data dump of the data of the error value identified or exceptional value is fallen, then use interpolation
Method replaces the value of missing, with supplement value mistake or the data of exception.Data for repeating record then can be eliminated directly.Its
In, whether repeat record refer to the identical record of attribute value in the database, identical can be examined by judging the record time
Survey whether record is identical, and identical record merges into a record, the elimination of the data of repetition record can be completed.
Further, it after completing data cleansing, also needs to be split data.For the segmentation of data can using with
Data after cleaning are equally divided into first part's data and second part data by machine mode, wherein first part's data are used for
Decision tree prediction model is trained, second part data are for verifying decision tree prediction model.
Step S30, based on treated data, decision tree prediction model is generated.
Specifically, as shown in figure 4, based on treated, data generation decision tree prediction model includes step in step S30
S301, step S302, step S303 and step S304.
Step S301, first part's data are pre-processed, obtains the input data and output of decision tree prediction model
Data.
Wherein, first part's data are pre-processed with the attribute mark specially by close window state in each data group
It is denoted as output, by other data group (outdoor temperature data group, outside relative humidity data group, outside atmospheric pressure data groups, room
External radiation intensity data group, outdoor air speed data group, outdoor wind direction data group, outdoor rainfall product data group attribute labeled as defeated
Enter), the input data and output data of decision tree prediction model can be obtained.That is the input data of decision tree prediction model includes
Outdoor temperature data group, outside relative humidity data group, outside atmospheric pressure data group, outdoor radiation intensity data group, outdoor
Air speed data group, outdoor wind direction data group, outdoor rainfall product data group, the output data of decision tree prediction model includes window
Switch state data group.
Step S302, input data and output data are imported, generating includes that multiple and different first decision trees for setting layers predict mould
Type.
It is to be appreciated that the process that decision tree generates is first to classify to input data, wherein assorting process uses base
In C4.5 algorithm, data group is classified at each node, to generate decision tree prediction model.Therefore, in step
In S302, specifically includes the following steps:
Step S3021, input data is handled based on C4.5 algorithm, obtains the letter of each feature in input data
Cease the ratio of gains.
Step S3022, the information gain ratio of each feature is compared judgement, obtains maximum first information gain
Than.
Step S3023, according to the maximum first information ratio of gains, spy corresponding with the maximum first information ratio of gains is determined
Sign.
Step S3024, using feature corresponding with the maximum first information ratio of gains as the decision of each node of decision tree
Point.
Step S3025, the information gain ratio of each characteristic value in decision point is calculated.
Step S3026, the information gain ratio of each characteristic value is compared judgement, obtains maximum second information and increases
Beneficial ratio.
Step S3027, it according to maximum second information gain ratio, determines with maximum second information gain than corresponding spy
Value indicative.
Step S3028, using with maximum second information gain than corresponding characteristic value as the further division of decision point
Cut-point.
Step S3029, input data is divided in the node based on decision point and cut-point.
Step S30220, it repeats the above steps, constantly divides each node, extend the tree layer depth of decision tree until default
The number of plies, to generate just decision tree prediction model.
Wherein, the maximum value of the first information ratio of gains is each layer of decision point (node i.e. above-mentioned).And the first letter
It is as follows to cease the maximum value solution procedure of the ratio of gains:
Wherein Info is the comentropy of entire data:
piFor the probability statistics situation of window state, p hereiFor percentage;I=1 is windowing;I=2 is to close window.
InfoTemperatureWhen to divide close window state using outdoor temperature, the comentropy of outdoor temperature:
pjFor grouping probability (such as outdoor temperature be 22-23 DEG C the case where the total situation of Zhan specific gravity, outdoor temperature 23-
24 DEG C of the case where the total situation of Zhan specific gravity, etc. and so on).Info (temperaturej) be under current state (such as outdoor temperature is
At 22-23 DEG C) the expectation value of information.When carrying out dividing close window state using the feature of other input datas, comentropy
Calculation method is similar.
After determining each decision point of decision tree, its cut-point is found out.Specifically, by taking outdoor temperature as an example, it is assumed that data
In the outdoor temperature data group collected in library, outdoor temperature minimum value is tminDEG C, maximum value tmaxDEG C, when cut-point is k DEG C
When, the node allocation at lower two child nodes the sum of comentropy:
Infok=Infot≥k℃+InfoT < k DEG C(tmin< k < tmax);
Then the second information gain is found than maximum value, and second information gain is than characteristic value corresponding to maximum value
For cut-point.
InfokWhen for temperature cut-point being k DEG C the node allocation at lower two child nodes the sum of comentropy,
Infot≥k℃For t >=k DEG C of comentropy, InfoT < k DEG CFor the comentropy of t < k DEG C.
Above is the assorting process of input data.Further, when some node output be same type (for example,
All it is that window is opened or window is closed), then stop classifying, and be ON or OFF by the node;If being not belonging to same type,
The above method is continued with to continue to divide.
In the continuous fission process of node, if when reach decision tree preset the number of plies when, even if node output be not all same
Type, still stops classification, and is the label of data that occupies the majority of output type (for example, 51% data by the vertex ticks
Output type be ON, the output types of remaining 49% data is OFF, then is ON by the vertex ticks).
It should be noted that the feature in input data is are as follows: outdoor temperature, outside relative humidity, outside atmospheric pressure,
Outdoor radiation intensity, outdoor wind speed, outdoor wind direction and outdoor rainfall, the characteristic value of each feature is the tool of this feature
Volume data, for example, some feature is outdoor temperature, then some characteristic value of this feature (outdoor temperature) is in outdoor temperature
Some temperature value.
As shown in figure 5, for the deep decision tree of different trees can be generated by setting the decision tree number of plies based on C4.5 algorithm
Prediction model.Wherein, Fig. 5 illustrates the decision tree prediction model schematic diagram that a depth is eight layers.
Step S303, according to the tree layer depth of decision tree and classification accuracy relationship, the layer of decision tree prediction model is determined
Number.
It is to be appreciated that the accuracy rate of prediction model classification can be improved by constantly increasing decision tree layer trial, however
The promotion that the decision tree number of plies attempts accuracy rate after increasing to a certain extent is no longer obvious, moreover, the too deep number of plies is but also model
Huge, model is complicated and easily leads to model prediction over-fitting.Therefore, it is necessary to carry out beta pruning to decision tree, determine that decision tree is predicted
The best number of plies of model.
As shown in fig. 6, the tree layer depth of decision tree and the relationship of classification accuracy are given, in figure, when the tree of decision tree
When layer depth is greater than 8, precision of prediction is no longer significantly increased, and therefore, the number of plies preferably used in example shown in Fig. 6 is 8.
Step S304, based on the identified number of plies, decision tree prediction model is generated.
Further, after generating decision tree prediction model using first part's data, then with second part data to mould
Type is checked and is verified, if model accuracy rate is not high, can adjust the prediction that relevant parameter improves model by trial-and-error method
Precision.Wherein it is possible to the parameter adjusted includes: the number of nodes of minimum number needed for classification, it is minimum required for a branch
Sample number, the smallest weight coefficient, maximum leaf segment points etc..
It should be noted that being related to a large amount of recursive calculations in decision tree prediction model generating process, therefore, specifically giving birth to
At can be realized in the process by software programs such as Matlab, Rapidminer, Python.
Step S40, according to decision tree prediction model, the state of output control signal control window.
After obtaining decision tree prediction model, by the way that the outdoor environment data at current time are imported decision tree prediction model
In, i.e., corresponding control signal is transmitted in the control equipment of window by predictable control of the indoor occupant to window in turn,
To control the state of window.
The present invention collects outdoor environment data by all kinds of environmental sensors in weather station, and combines indoor occupant to window
Data are controlled, establish close window model to predict and imitate indoor occupant and open a window/close window behavior, it is automatic that window control is driven to set
It is standby, to control the open state of window, to realize in the case where unmanned manipulation, automatically according to indoor occupant demand, introduce
Natural wind improves indoor thermal environment, and reduces air conditioning energy consumption.
Further, as shown in fig. 7, the present invention also provides the window control device based on decision tree prediction model, packet
It includes: acquiring unit 701, processing unit 702, execution unit 703 and control unit 704.
Acquiring unit 701, for obtaining indoor occupant to the control data and environmental data of window;Environmental data includes room
External environment data.
Processing unit 702, for handling control data and outdoor environment data, to obtain treated data.
Execution unit 703, for generating decision tree prediction model based on treated data.
Control unit 704, for according to decision tree prediction model, the state of output control signal control window.
It is to be appreciated that being somebody's turn to do the window control device based on decision tree prediction model can be used to implement the embodiment of the present invention
The provided window control method based on decision tree prediction model.
As shown in figure 8, the present invention also provides a kind of electronic equipment, including memory and processor, processor are used for basis
The program instruction that memory is stored executes the step of window control method based on decision tree prediction model.
The present invention also provides a kind of storage mediums, are stored thereon with computer program, and computer program is executed by processor
Shi Shixian based on decision tree prediction model window control method the step of.
Above embodiments only technical concepts and features to illustrate the invention, its object is to allow person skilled in the art
Scholar can understand the contents of the present invention and implement accordingly, can not limit the scope of the invention.It is all to be wanted with right of the present invention
The equivalent changes and modifications that range is done are sought, should belong to the covering scope of the claims in the present invention.
It should be understood that for those of ordinary skills, it can be modified or changed according to the above description,
And all these modifications and variations should all belong to the protection domain of appended claims of the present invention.
Claims (11)
1. a kind of window control method based on decision tree prediction model characterized by comprising
Indoor occupant is obtained to the control data and environmental data of window;The environmental data includes outdoor environment data;
The control data and the outdoor environment data are handled, to obtain treated data;
Based on treated the data, decision tree prediction model is generated;
According to the decision tree prediction model, the state of output control signal control window.
2. the window control method according to claim 1 based on decision tree prediction model, which is characterized in that the interior
Personnel include the control data bag of window: window turn-on data and closing data;
The location information of the window is obtained by photoelectric sensor or position sensor detection;
The outdoor environment data include: outdoor temperature, outside relative humidity, outside atmospheric pressure, outdoor radiation intensity, outdoor
Wind speed, outdoor wind direction and outdoor rainfall.
3. the window control method according to claim 1 based on decision tree prediction model, which is characterized in that described to institute
It states control data and the outdoor environment data is handled, treated that data include: to obtain
With the acquisition time of the control data and outdoor environment data sequence, to the control data and the outdoor ring
Border data are arranged;
By after arrangement control data and outdoor environment data be stored in database profession.
4. the window control method according to claim 3 based on decision tree prediction model, which is characterized in that will arrange
After control data and outdoor environment data afterwards are stored in database profession further include:
The data being stored in database profession are cleaned, with the data after being cleaned;
Data after cleaning are split, to obtain the segmentation data for generating the decision tree prediction model;Described point
Cutting data is treated the data.
5. the window control method according to claim 4 based on decision tree prediction model, which is characterized in that described pair is deposited
The data being stored in database carry out cleaning
The data of value missing are found out, and replace the value of missing, the data lacked with supplement value using interpolation method;
Using method of analysis of variance, value mistake or abnormal data are identified;
The data of described value mistake or exception are removed, and replace the value of missing using interpolation method, to supplement described value mistake or different
Normal data;
Eliminate the data for repeating record.
6. the window control method according to claim 4 based on decision tree prediction model, which is characterized in that described to clear
Data after washing are split, and the segmentation data with acquisition for generating the decision tree prediction model include:
Data after cleaning are equally divided by first part's data and second part data using random fashion, wherein described
A part of data are for being trained the decision tree prediction model, and the second part data are for pre- to the decision tree
Model is surveyed to be verified.
7. the window control method according to claim 6 based on decision tree prediction model, which is characterized in that described to be based on
Treated the data, generating decision tree prediction model includes:
First part's data are pre-processed, the input data and output data of decision tree prediction model are obtained;It is described
Input data includes: outdoor temperature, outside relative humidity, outside atmospheric pressure, outdoor radiation intensity, outdoor wind speed, outdoor wind
To and outdoor rainfall;The output data is the state of window;
The input data and output data are imported, generating includes multiple and different first decision tree prediction models for setting layers;
According to the tree layer depth of decision tree and classification accuracy relationship, the number of plies of the decision tree prediction model is determined;
Based on the identified number of plies, the decision tree prediction model is generated.
8. the window control method according to claim 7 based on decision tree prediction model, which is characterized in that the importing
The input data and output data, generating includes that multiple and different first decision tree prediction models for setting layers include:
The input data is handled based on C4.5 algorithm, obtains the information gain of each feature in the input data
Than;
The information gain ratio of each feature is compared judgement, obtains the maximum first information ratio of gains;
According to the maximum first information ratio of gains, determining feature corresponding with the maximum first information ratio of gains;
Using feature corresponding with the maximum first information ratio of gains as the decision point of each node of the decision tree;
Calculate the information gain ratio of each characteristic value in the decision point;
The information gain ratio of each characteristic value is compared judgement, obtains maximum second information gain ratio;
According to the maximum second information gain ratio, determine with maximum second information gain than corresponding characteristic value;
Using with maximum second information gain than corresponding characteristic value as the segmentation of the decision point further divided
Point;
Input data is divided in the node based on the decision point and the cut-point;
It repeats the above steps, constantly divides each node, extend the tree layer depth of decision tree until presetting the number of plies, described in generating
First decision tree prediction model.
9. a kind of window control device based on decision tree prediction model characterized by comprising
Acquiring unit, for obtaining indoor occupant to the control data and environmental data of window;The environmental data includes outdoor
Environmental data;
Processing unit, for handling the control data and the outdoor environment data, to obtain treated data;
Execution unit, for generating decision tree prediction model based on treated the data;
Control unit, for according to the decision tree prediction model, the state of output control signal control window.
10. a kind of electronic equipment, which is characterized in that including memory and processor, the processor is used for according to the storage
The program instruction perform claim that device is stored requires the step of any one of 1-8 the method.
11. a kind of storage medium, is stored thereon with computer program, which is characterized in that the computer program is held by processor
It is realized when row such as the step of any one of claim 1-8 the method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910492486.8A CN110409955A (en) | 2019-06-06 | 2019-06-06 | Window control method, device and electronic equipment based on decision tree prediction model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910492486.8A CN110409955A (en) | 2019-06-06 | 2019-06-06 | Window control method, device and electronic equipment based on decision tree prediction model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110409955A true CN110409955A (en) | 2019-11-05 |
Family
ID=68358429
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910492486.8A Pending CN110409955A (en) | 2019-06-06 | 2019-06-06 | Window control method, device and electronic equipment based on decision tree prediction model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110409955A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116256980A (en) * | 2023-05-15 | 2023-06-13 | 四川名人居门窗有限公司 | Indoor environment self-learning control method and system |
CN116805228A (en) * | 2023-08-18 | 2023-09-26 | 鲲鹏智慧冷链(山东)有限公司 | Cold-chain logistics storage temperature and humidity prediction system and method |
CN117823010A (en) * | 2023-12-26 | 2024-04-05 | 湖北工程学院新技术学院 | Intelligent window opening and closing system and method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3046290A1 (en) * | 2015-01-15 | 2016-07-20 | Xiaomi Inc. | Method and device for generating prompt |
CN105955356A (en) * | 2016-05-05 | 2016-09-21 | 华南理工大学 | Indoor environment control system and method based on human body heat acclimatization |
CN106223771A (en) * | 2016-08-26 | 2016-12-14 | 宁波博报门窗有限公司 | A kind of smart window and control method thereof |
CN107561997A (en) * | 2017-08-22 | 2018-01-09 | 电子科技大学 | A kind of power equipment state monitoring method based on big data decision tree |
JP2018028249A (en) * | 2016-08-19 | 2018-02-22 | 文化シヤッター株式会社 | Opening/closing control system and opening/closing control method |
CN108999517A (en) * | 2018-08-06 | 2018-12-14 | 深圳汇创联合自动化控制有限公司 | Smart home Ventilation conditioning system |
CN109451751A (en) * | 2018-05-31 | 2019-03-08 | 深圳市蚂蚁雄兵物联技术有限公司 | Storage medium and home equipment intelligent controlling device, method and system |
-
2019
- 2019-06-06 CN CN201910492486.8A patent/CN110409955A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP3046290A1 (en) * | 2015-01-15 | 2016-07-20 | Xiaomi Inc. | Method and device for generating prompt |
CN105955356A (en) * | 2016-05-05 | 2016-09-21 | 华南理工大学 | Indoor environment control system and method based on human body heat acclimatization |
JP2018028249A (en) * | 2016-08-19 | 2018-02-22 | 文化シヤッター株式会社 | Opening/closing control system and opening/closing control method |
CN106223771A (en) * | 2016-08-26 | 2016-12-14 | 宁波博报门窗有限公司 | A kind of smart window and control method thereof |
CN107561997A (en) * | 2017-08-22 | 2018-01-09 | 电子科技大学 | A kind of power equipment state monitoring method based on big data decision tree |
CN109451751A (en) * | 2018-05-31 | 2019-03-08 | 深圳市蚂蚁雄兵物联技术有限公司 | Storage medium and home equipment intelligent controlling device, method and system |
CN108999517A (en) * | 2018-08-06 | 2018-12-14 | 深圳汇创联合自动化控制有限公司 | Smart home Ventilation conditioning system |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116256980A (en) * | 2023-05-15 | 2023-06-13 | 四川名人居门窗有限公司 | Indoor environment self-learning control method and system |
CN116256980B (en) * | 2023-05-15 | 2023-07-25 | 四川名人居门窗有限公司 | Indoor environment self-learning control method and system |
CN116805228A (en) * | 2023-08-18 | 2023-09-26 | 鲲鹏智慧冷链(山东)有限公司 | Cold-chain logistics storage temperature and humidity prediction system and method |
CN117823010A (en) * | 2023-12-26 | 2024-04-05 | 湖北工程学院新技术学院 | Intelligent window opening and closing system and method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
KR102159692B1 (en) | solar photovoltatic power generation forecasting apparatus and method based on big data analysis | |
CN110409955A (en) | Window control method, device and electronic equipment based on decision tree prediction model | |
CN108733966A (en) | A kind of multidimensional electric energy meter field thermodynamic state verification method based on decision woodlot | |
CN106952028A (en) | Dynamoelectric equipment failure is examined and health control method and system in advance | |
CN109858569A (en) | Multi-tag object detecting method, system, device based on target detection network | |
CN109146162B (en) | A kind of probability wind speed forecasting method based on integrated Recognition with Recurrent Neural Network | |
CN107909211B (en) | Wind field equivalent modeling and optimization control method based on fuzzy c-means clustering algorithm | |
CN106502871B (en) | The alarm threshold dynamic configuration system and method for supervisory systems | |
CN103853106A (en) | Energy consumption prediction parameter optimization method of building energy supply device | |
CN108090628A (en) | A kind of grain feelings security detection and analysis method based on PSO-LSSVM algorithms | |
CN106021771A (en) | Method and device for diagnosing faults | |
CN110020712B (en) | Optimized particle swarm BP network prediction method and system based on clustering | |
CN110264116A (en) | A kind of Electrical Power System Dynamic safety evaluation method explored based on relationship with regression tree | |
CN111680820A (en) | Distributed photovoltaic power station fault diagnosis method and device | |
CN110119766A (en) | A kind of multiple groups close the green pepper greenhouse temperature intelligence prior-warning device of intelligent model | |
CN116316599A (en) | Intelligent electricity load prediction method | |
Kadir et al. | Wheat yield prediction: Artificial neural network based approach | |
CN108960530A (en) | Prediction technique based on the long crop field vegetation coverage index of memory network in short-term | |
CN107219759A (en) | A kind of Technique for Controlling Greenhouse Environment and device | |
CN116799796A (en) | Photovoltaic power generation power prediction method, device, equipment and medium | |
WO2022039675A1 (en) | Method and apparatus for forecasting weather, electronic device and storage medium thereof | |
Culclasure | Using neural networks to provide local weather forecasts | |
CN110489852A (en) | Improve the method and device of the wind power system quality of data | |
CN107976919A (en) | A kind of Study of Intelligent Robot Control method, system and electronic equipment | |
Lu et al. | Image classification and identification for rice leaf diseases based on improved WOACW_SimpleNet |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191105 |
|
RJ01 | Rejection of invention patent application after publication |