CN111798066A - Multi-dimensional prediction method and system for cell flow under urban scale - Google Patents
Multi-dimensional prediction method and system for cell flow under urban scale Download PDFInfo
- Publication number
- CN111798066A CN111798066A CN202010693644.9A CN202010693644A CN111798066A CN 111798066 A CN111798066 A CN 111798066A CN 202010693644 A CN202010693644 A CN 202010693644A CN 111798066 A CN111798066 A CN 111798066A
- Authority
- CN
- China
- Prior art keywords
- prediction
- model
- data
- time
- different
- 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
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000012549 training Methods 0.000 claims abstract description 46
- 238000003860 storage Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000013467 fragmentation Methods 0.000 claims description 2
- 238000006062 fragmentation reaction Methods 0.000 claims description 2
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 abstract 1
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 10
- 238000004590 computer program Methods 0.000 description 7
- 230000008859 change Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000009499 grossing Methods 0.000 description 4
- 230000015654 memory Effects 0.000 description 4
- 230000001413 cellular effect Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 241001123248 Arma Species 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000012795 verification Methods 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"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/60—Business processes related to postal services
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Marketing (AREA)
- Biophysics (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Tourism & Hospitality (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The utility model provides a multi-dimensional prediction method and system for regional flow under urban scale, which collects flow data in the region and preprocesses the collected data; different prediction models are constructed in advance, the prediction models are trained, and the prediction period and the moving time of training data of each prediction model are different; and (4) carrying out statistics by taking the error rate and the filling rate as indexes, evaluating the prediction accuracy and performance of different models in different dimensions, and selecting an optimized model to perform flow prediction to obtain a prediction result. The artificial participation degree is reduced, the operation efficiency is improved, and the convergence of the prediction model is better when the MAPE value is verified to be lower, so that the prediction accuracy is improved, an independent prediction model is trained by more than 5 ten thousand network elements, and the accurate prediction of each network element is realized.
Description
Technical Field
The disclosure belongs to the technical field of pattern recognition, and relates to a multi-dimensional prediction method and a multi-dimensional prediction system for cell flow under an urban scale.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
With the continuous development of human society, cities below 5G will bear more and more intelligent terminal devices. The research of community traffic has important research significance on all social levels, such as the field of intelligent cities, the deployment and distribution of 5G network regional network resources and the number of base stations from the global perspective are realized through traffic prediction, such as planning geographic positions and the like, on the premise of ensuring the traffic demand of a community, and whether more or fewer base stations are opened is determined according to the predicted service load. The prediction technology can also be applied to the aspects of network capacity prediction, network performance early warning, user perception guarantee and the like.
With the continuous development of mobile communication technology, mobile cellular networks are applied more and more widely. The traffic of the cellular network generates a large amount of data and consumes a large amount of power. In order to satisfy the competitive design tradeoff between data service bandwidth and hardware energy cost, it is important to analyze and predict service data of a cellular network base station with emphasis while satisfying the quality of service (QoS) requirements, and to reduce the energy consumption of the base station. In particular, a base station is generally composed of three cells. When the upcoming data traffic is predicted to be low, the whole base station or a specific cell of the base station is closed in advance to reduce energy consumption; when the increase of the traffic is predicted, the base station is opened in advance to ensure the data QoS. Traffic prediction is not only a problem that an operator knows in advance to consider service perception of a user, but also a problem that the operator models overall distribution of data, groups multidimensional space and time, and predicts effectiveness of a plurality of scenes to predict application conditions of the user in the scenes.
"5 +5+ 4" multidimensional prediction. Wherein the time dimension is from year, month, week, day and hour from large to small; the spatial dimension is five levels from big to small, namely, a full network level, a district level, a county level, a micro-area level, a grid level and a cell level; the independent dimension is four types including a manufacturer, a coverage scene, a coverage type and a frequency band, and all the dimensions except the dimension of the manufacturer can be drilled below a cell level. There are a total of 5 tens of thousands of network elements that need to be predicted.
Various network capacity and flow prediction methods are provided aiming at different problems, and the methods comprise smoothing, trend fitting, a combined model, an AR model, an MA model, an ARMA model and an ARIMA model. Generally, the above method is to find the most suitable predicted trend line according to the daily traffic history data of the whole network. The accuracy of these methods is limited by factors such as historical data integrity, size of the data volume, etc. The prediction result often has an over-fitting or under-fitting phenomenon, and the deviation of the prediction result and actual data is often not obvious. With the gradual maturity of machine learning-based methods, the deep learning field is continuously developed and has been applied to traffic prediction in different scenes.
The inventor finds that an exponential smoothing method and a trend fitting method are common methods for early time series prediction, but the exponential smoothing method and the trend fitting method are not suitable for frequently changing data sets, cannot process nonlinear time series, and can only fit results of complex change rules by one method; the prediction of the AR/MA/ARMA/ARIMA model requires a plurality of manual adjustment parameters, and has low efficiency and inaccurate prediction.
Disclosure of Invention
The method and the system can reduce the manual workload, and are high in efficiency and high in prediction accuracy.
According to some embodiments, the following technical scheme is adopted in the disclosure:
a multi-dimensional prediction method for regional flow under an urban scale comprises the following steps:
collecting flow data in the area, and preprocessing the collected data;
different prediction models are constructed in advance, the prediction models are trained, and the prediction period and the moving time of training data of each prediction model are different;
and (4) carrying out statistics by taking the error rate and the filling rate as indexes, evaluating the prediction accuracy and performance of different models in different dimensions, and selecting an optimized model to perform flow prediction to obtain a prediction result.
As an alternative embodiment, the predictive model comprises at least seven of:
model 1: one week after each prediction, the training data is shifted one day each time;
model 2: one week after each prediction, the training data is shifted one week each time;
model 3: two weeks after each prediction, the training data was shifted one day each time;
model 4: two weeks after each prediction, the training data was shifted one week each time;
model 5: four weeks after each prediction, the training data is shifted one day each time;
model 6: four weeks after each prediction, the training data is shifted one week each time;
model 7: one day after each prediction, the training data was shifted one day each.
As an alternative, when the prediction accuracy and performance of different models in different dimensions are evaluated, the selection of the optimized model is performed before each network element predicts.
As an alternative embodiment, before prediction, the regional traffic data is collected and normalized, and the regional traffic in the whole network is mapped into the range of [0,1 ].
As an alternative embodiment, the collected traffic data is divided into a training set and a test set.
As an alternative embodiment, the data pre-processing process includes backfilling, deduplication and fragmentation of the acquired data.
As an alternative embodiment, after the prediction is completed, the predicted value is restored to the traffic data with the time stamp by date before the data normalization.
As an alternative embodiment, the prediction granularity is composed of three dimensions of time, space and independence, and each dimension has a plurality of layers.
As an alternative embodiment, the optimization model is an optimal prediction model combining temporal granularity and spatial granularity.
A multi-dimensional prediction system for regional traffic under a city scale comprises:
the data acquisition and processing module is configured to acquire flow data in the region and preprocess the acquired data;
the model training module is configured to train different pre-constructed prediction models, and the prediction period and the moving time of training data of each prediction model are different;
and the model evaluation and selection module is configured to evaluate the prediction accuracy and performance of different models in different dimensions by taking the error rate and the filling rate as index statistics, and select the optimized model to perform flow prediction to obtain a prediction result.
A computer readable storage medium having stored therein a plurality of instructions adapted to be loaded by a processor of a terminal device and to execute the method for multi-dimensional prediction of regional traffic at a city scale.
A terminal device comprising a processor and a computer readable storage medium, the processor being configured to implement instructions; the computer readable storage medium is used for storing a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the regional flow multidimensional prediction method under the urban scale.
Compared with the prior art, the beneficial effect of this disclosure is:
according to the method, the time characteristics and the behavior rules of the model data are deeply researched through statistical analysis of the urban scale real traffic data with the time span of more months. Considering that the time sequence data show different change rules under different time dimensions, space dimensions and independent dimensions, a multi-dimensional recurrent neural network (MDRNN) prediction model is established, and the future cell telephone traffic under different time dimensions and space dimensions is predicted.
The method fully utilizes training and inspection of actual data, the granularity of the provided prediction model can be deep to a unit level, compared with the traditional trend fitting method, the Mean Absolute Percentage Error (MAPE) of the model is reduced by 6.56%, and guidance is provided for energy efficiency optimization and power consumption reduction of a base station in different space-time dimensions. The manual workload is reduced, the efficiency is high, and the prediction accuracy is high.
The method and the device do not need to manually appoint the cycle of the time sequence or parameters such as difference, the manual participation degree is reduced, the operation efficiency is improved, the convergence is better, the prediction accuracy is improved, the independent prediction model is trained by more than 5 ten thousand network elements, and the accurate prediction of each network element is realized.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.
FIG. 1 is a flow diagram of a flow prediction for a multi-dimensional prediction model;
fig. 2 is an overall framework diagram of the automatic flow prediction system.
The specific implementation mode is as follows:
the present disclosure is further described with reference to the following drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
First, the term explanation is made:
(1) data normalization: the goal is to map the full network cell traffic into the range of [0,1 ]. Normalization refers to scaling the data to fall within a small specific interval. Since RNN is sensitive to the size of the input data, it is necessary to map the data uniformly into the range of [0,1 ]. After normalization, the convergence rate of the model and the precision of the model are improved.
(2) And (3) data shape conversion: 2D: [ samples, features ] to 3D: [ samples, timepieces, features ] conversion. Since the RNN network expects the incoming data to be 3D, Timesteps is set to 1 when converting from 2-dimensional to 3-dimensional.
(3) Data set partitioning: the data set is divided into a training set (train _ data) and a test set (test _ data). Training data: the test data is about 7.5: 1.
(4) and (3) a time sequence prediction method: according to the time sequence of historical statistical data, the future change trend is subjected to predictive analysis. In general, a time series consists of four variation components, such as long-term trend variations, seasonal variations, periodic variations, and random fluctuations. Some simple prediction models can be used to predict the three trend changes, such as an exponential smoothing model, a moving average model, and the like. The random variation component is unpredictable, and is a kind of "noise" mixed into the time series, and must be filtered out in an attempt to avoid affecting the accuracy of the prediction result.
(5) Rnn (current Neural networks): the recurrent neural network is a neural network with memory and can remember the things at the previous time point. It can be treated as a black box that can fit any function, as long as the training data is sufficient to obtain the desired output given a particular input.
(6) LSTM (Long Short-Term Memory): is a specific RNN that remembers longer memory relative to RNNs, but is still short-term in nature. The LSTM model is more complex than the RNN model, and training data needs to be more time consuming than traditional RNNs.
The traditional method is limited by factors such as the integrity of historical data, the size of a data gauge and the like, the prediction result is often over-fit or under-fit, the prediction result is compared with real data, the deviation is often large, and the prediction cannot be carried out to the cell level.
In order to solve the above problems, a first aspect of the present invention provides a prediction method using a multidimensional prediction model, which reduces the workload of workers, and has high efficiency and high prediction accuracy. As shown in fig. 1, the method for constructing a multidimensional prediction model based on RNN, Python, and Oracle in this embodiment includes:
(1) saving models after each iteration
In the training process, the training is carried out according to the cell data files under the catalog, so the trained model of the cell is saved according to the cell name each time, the cell name is stored by ci and saved by using cover.
(2) Selecting an optimization model
The output _ mse is a dictionary type, the stored content is ({ (step length: corresponding loss value), () }), and the key { (output _ mse. get) obtains the step length corresponding to the minimum loss value.
(3) Saving optimal models
According to the iteration step length corresponding to the obtained optimal model, the optimal model is saved by a code "save _ path [ min _ ep ] ═ save (sess, save _ path _1+ '/model/' + str (ci) +". ckpt ")" so as to be directly called and used later, and thus, multiple times of training in the same cell can be reduced.
(4) Loading the saved optimal model
And carrying out data loading on the optimal model stored in the last step.
(5) Predicting data
Outputs refers to the predicted output values and feed _ fact refers to the input data.
(6) Restoring the predicted value to the state before normalization
And restoring the predicted value into the flow data of the hour level with the time-of-day as the time stamp before the data normalization.
The different models formed include: model 1: one week after each prediction, the training data was shifted one day each, model 2: one week after each prediction, one week for each shift of training data, model 3: two weeks after each prediction, the training data was shifted one day each, model 4: two weeks after each prediction, the training data was shifted one week each, model 5: four weeks after each prediction, the training data was shifted one day each time, model 6: four weeks after each prediction, the training data moved one week each time, model 7: one day after each prediction, the training data was shifted one day each. Each network element needs to make an evaluation and automatic selection of 7 models.
In this embodiment, the prediction granularity is composed of three dimensions of time, space and independence, and each dimension has a plurality of levels. The spatial dimension is five levels from big to small, namely, a full network level, a district level, a county level, a micro-area level, a grid level and a cell level; the independent dimensionalities comprise four types, namely a manufacturer, a coverage scene, a coverage type and a frequency band, and all the dimensionalities except the manufacturer dimensionality can be drilled down to a cell level; the time dimension is five levels from large to small, namely year, month, week, day and hour.
Due to the use of RNN for prediction, enough data is needed to represent a better advantage. The change rule of 24-hour traffic of each cell in one day can be better reflected by selecting hour-level data for prediction, and the time period in which the cell is not busy can be predicted. However, some cells have deletions in the data aggregation process, so that the daily data fluctuation is large, and the trend of the daily overall traffic is influenced. Therefore, in the process of predicting by using the hour data, the hour data is preprocessed, for example, to reflect the change rule of one week, so when the missing value of the hour data is filled, a weighted average is used to fill the missing value of the hour data by using the data at the same time point in the previous week and the next week.
The result of the embodiment is an optimal model, the optimal prediction model combining the time granularity and the space granularity is given by the model, and the prediction accuracy of the model is verified to be high by testing on an independent dimension.
In this embodiment, for a plurality of network elements of 5 ten thousand, a model after each iteration of each model can be automatically stored, and the prediction accuracy and performance of the model in different dimensions are evaluated by taking the error rate and the fill factor as indexes from the automatically stored models. The filling rate is low, the error rate is the minimum optimal model, the system automatically selects the optimal model during training, the optimal model during training is directly used during prediction, and the flow prediction is carried out after the optimal model is loaded.
Fig. 2 shows the system implementation process of data acquisition → data preprocessing → model auto-training → data prediction → data output and storage Shell.
(1) Data acquisition: an Oracle database is used for design, and a design database storage process is developed and used for timing automatic acquisition of flow data. This phase is expected to take 10 hours.
(2) Data preprocessing: the method comprises the steps of adopting Java language, exporting collected data to a server in a text mode, designing and developing a data preprocessing module, and carrying out 3 preprocessing operations of backfilling, removing duplicate and fragmenting on the collected data (as the RNN needs continuous data and cannot have abnormal values when carrying out model training, a program is needed to automatically preprocess the data). Data pretreatment for this period was expected to be 10 minutes.
(3) Automatic training of a model: and respectively carrying out model training on the 5+5+4 dimensionality fragment original data by adopting a Python language to realize a special model for each network element, and storing model parameters into a model result set.
(4) And (3) data prediction: and adopting Python language, selecting an optimal model from the model result set, respectively predicting each network element for more than 5 ten thousand times in total, and storing the prediction result into a text. Data training and prediction two phases are expected to take 3 hours.
(5) Data output and storage: and (3) outputting and storing the predicted data: and developing a shell script, and warehousing the prediction result into an Oracle database, so that result verification and prediction result query are facilitated. One round of model training and prediction is expected to be completed within 24 hours.
In order to reduce the manual workload and improve the efficiency, the embodiment utilizes the model to automatically train and predict, and puts the prediction result into the Oracle database, thereby reducing the manual participation, reducing the manual workload and improving the efficiency.
As will be appreciated by one skilled in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and so forth) having computer-usable program code embodied therein.
The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
Although the present disclosure has been described with reference to specific embodiments, it should be understood that the scope of the present disclosure is not limited thereto, and those skilled in the art will appreciate that various modifications and changes can be made without departing from the spirit and scope of the present disclosure.
Claims (10)
1. A multi-dimensional prediction method for regional flow under an urban scale is characterized by comprising the following steps: the method comprises the following steps:
collecting flow data in the area, and preprocessing the collected data;
different prediction models are constructed in advance, the prediction models are trained, and the prediction period and the moving time of training data of each prediction model are different;
and (4) carrying out statistics by taking the error rate and the filling rate as indexes, evaluating the prediction accuracy and performance of different models in different dimensions, and selecting an optimized model to perform flow prediction to obtain a prediction result.
2. The method for predicting the regional flow under the urban scale in multiple dimensions as claimed in claim 1, wherein the method comprises the following steps: the predictive model includes at least seven types:
model 1: one week after each prediction, the training data is shifted one day each time;
model 2: one week after each prediction, the training data is shifted one week each time;
model 3: two weeks after each prediction, the training data was shifted one day each time;
model 4: two weeks after each prediction, the training data was shifted one week each time;
model 5: four weeks after each prediction, the training data is shifted one day each time;
model 6: four weeks after each prediction, the training data is shifted one week each time;
model 7: one day after each prediction, the training data was shifted one day each.
3. The method for predicting the regional flow under the urban scale in multiple dimensions as claimed in claim 1, wherein the method comprises the following steps:
before prediction, the regional flow data is collected and normalized, and the flow in the whole network region is mapped into the range of [0,1 ].
4. The method for predicting the regional flow under the urban scale in multiple dimensions as claimed in claim 1, wherein the method comprises the following steps: dividing the collected flow data into a training set and a testing set;
the data preprocessing process comprises backfilling, duplicate removal and fragmentation of the collected data.
5. The method for predicting the regional flow under the urban scale in multiple dimensions as claimed in claim 1, wherein the method comprises the following steps: and after the prediction is finished, restoring the predicted value into the flow data with the time stamp by date and time before the data normalization.
6. The method for predicting the regional flow under the urban scale in multiple dimensions as claimed in claim 1, wherein the method comprises the following steps: the prediction granularity is composed of three dimensions of time, space and independence, and each dimension has a plurality of layers.
7. The method for predicting the regional flow under the urban scale in multiple dimensions as claimed in claim 6, wherein the method comprises the following steps: the optimization model is an optimal prediction model combining time granularity and space granularity.
8. A multi-dimensional prediction system for regional flow under urban scale is characterized in that: the method comprises the following steps:
the data acquisition and processing module is configured to acquire flow data in the region and preprocess the acquired data;
the model training module is configured to train different pre-constructed prediction models, and the prediction period and the moving time of training data of each prediction model are different;
and the model evaluation and selection module is configured to evaluate the prediction accuracy and performance of different models in different dimensions by taking the error rate and the filling rate as index statistics, and select the optimized model to perform flow prediction to obtain a prediction result.
9. A computer-readable storage medium characterized by: a plurality of instructions are stored, the instructions are suitable for being loaded by a processor of a terminal device and executing the regional flow multidimensional prediction method under the urban scale according to any one of claims 1-7.
10. A terminal device is characterized in that: the system comprises a processor and a computer readable storage medium, wherein the processor is used for realizing instructions; the computer-readable storage medium is used for storing a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the regional flow multidimensional prediction method under the urban scale according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010693644.9A CN111798066A (en) | 2020-07-17 | 2020-07-17 | Multi-dimensional prediction method and system for cell flow under urban scale |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010693644.9A CN111798066A (en) | 2020-07-17 | 2020-07-17 | Multi-dimensional prediction method and system for cell flow under urban scale |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111798066A true CN111798066A (en) | 2020-10-20 |
Family
ID=72807750
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010693644.9A Pending CN111798066A (en) | 2020-07-17 | 2020-07-17 | Multi-dimensional prediction method and system for cell flow under urban scale |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111798066A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113988452A (en) * | 2021-11-08 | 2022-01-28 | 成都四方伟业软件股份有限公司 | Network element alarm prediction method and device based on stacked LSTM |
| CN115225518A (en) * | 2021-03-29 | 2022-10-21 | 中国移动通信集团福建有限公司 | Base station traffic processing method and device and network equipment |
| CN116504076A (en) * | 2023-06-19 | 2023-07-28 | 贵州宏信达高新科技有限责任公司 | Expressway traffic flow prediction method based on ETC portal data |
-
2020
- 2020-07-17 CN CN202010693644.9A patent/CN111798066A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115225518A (en) * | 2021-03-29 | 2022-10-21 | 中国移动通信集团福建有限公司 | Base station traffic processing method and device and network equipment |
| CN115225518B (en) * | 2021-03-29 | 2024-04-12 | 中国移动通信集团福建有限公司 | Base station traffic processing method and device and network equipment |
| CN113988452A (en) * | 2021-11-08 | 2022-01-28 | 成都四方伟业软件股份有限公司 | Network element alarm prediction method and device based on stacked LSTM |
| CN116504076A (en) * | 2023-06-19 | 2023-07-28 | 贵州宏信达高新科技有限责任公司 | Expressway traffic flow prediction method based on ETC portal data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11876374B2 (en) | System and method for optimal control of energy storage system | |
| CN111798066A (en) | Multi-dimensional prediction method and system for cell flow under urban scale | |
| CN101442762B (en) | Method and apparatus for analyzing network performance and locating network fault | |
| CN110222882A (en) | A kind of prediction technique and device of electric system Mid-long Term Load | |
| CN109587713A (en) | A kind of network index prediction technique, device and storage medium based on ARIMA model | |
| CN111680841B (en) | Short-term load prediction method, system and terminal equipment based on principal component analysis | |
| CN116993227B (en) | Heat supply analysis and evaluation method, system and storage medium based on artificial intelligence | |
| KR102518629B1 (en) | Method for forecasting electric power demand using convolutional neural network, recording medium and device for performing the method | |
| CN103425564B (en) | A kind of smartphone software uses Forecasting Methodology | |
| CN103745280A (en) | Prediction method, device and processor for electricity consumption | |
| CN112163723B (en) | Hydropower station medium-long-term runoff prediction method, medium and equipment based on scene division | |
| CN114330934A (en) | Model parameter self-adaptive GRU new energy short-term power generation power prediction method | |
| CN113570132A (en) | Wind power prediction method for space-time meteorological feature extraction and deep learning | |
| CN102065449A (en) | Method for predicting mobile communication telephone traffic based on clustered LS-SVM (Least Squares-Support Vector Machine) | |
| CN117974364A (en) | Emission and energy consumption data acquisition and metering method and system based on electric power system | |
| CN118134141A (en) | Energy distribution method and system for distributed energy storage micro-energy network | |
| CN110909916B (en) | Wind power generation month electric quantity interval prediction method based on entropy weight method | |
| Mashlakov et al. | Multi-timescale forecasting of battery energy storage state-of-charge under frequency containment reserve for normal operation | |
| CN115495862A (en) | Power transmission network extension planning method and system considering extreme scenes of renewable energy sources | |
| CN114372615A (en) | Short-term photovoltaic power prediction method and system considering spatial correlation | |
| CN108615091A (en) | Electric power meteorology load data prediction technique based on cluster screening and neural network | |
| Zhao et al. | Forecasting regional short-term freight volume using QPSO-LSTM algorithm from the perspective of the importance of spatial information | |
| CN114091325A (en) | Power distribution station load prediction method and system based on multi-model layered learning | |
| Bhatnagar et al. | Comprehensive Electric load forecasting using ensemble machine learning methods | |
| Yanhui et al. | Forecasting hourly electricity demand with nonparametric functional data analysis |
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 |