CN117112859B - Display method, device and computer readable storage medium for population movement evolution - Google Patents
Display method, device and computer readable storage medium for population movement evolution Download PDFInfo
- Publication number
- CN117112859B CN117112859B CN202310730446.9A CN202310730446A CN117112859B CN 117112859 B CN117112859 B CN 117112859B CN 202310730446 A CN202310730446 A CN 202310730446A CN 117112859 B CN117112859 B CN 117112859B
- Authority
- CN
- China
- Prior art keywords
- population
- migration
- data
- features
- image
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000003860 storage Methods 0.000 title claims abstract description 9
- 238000013508 migration Methods 0.000 claims abstract description 128
- 230000005012 migration Effects 0.000 claims abstract description 128
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 79
- 238000007637 random forest analysis Methods 0.000 claims abstract description 59
- 230000000007 visual effect Effects 0.000 claims abstract description 27
- 230000008569 process Effects 0.000 claims abstract description 23
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000013507 mapping Methods 0.000 claims abstract description 7
- 238000010295 mobile communication Methods 0.000 claims description 14
- 238000012216 screening Methods 0.000 claims description 12
- 238000001914 filtration Methods 0.000 claims description 11
- 238000013519 translation Methods 0.000 claims description 10
- 230000003993 interaction Effects 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims 1
- 238000004364 calculation method Methods 0.000 abstract description 43
- 238000003066 decision tree Methods 0.000 description 67
- 238000000605 extraction Methods 0.000 description 12
- 238000012545 processing Methods 0.000 description 12
- 230000009467 reduction Effects 0.000 description 12
- 238000012549 training Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 9
- 238000005065 mining Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 230000002159 abnormal effect Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 4
- 230000010365 information processing Effects 0.000 description 4
- 238000003058 natural language processing Methods 0.000 description 4
- 238000007619 statistical method Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 238000009472 formulation Methods 0.000 description 3
- 238000009499 grossing Methods 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 239000000203 mixture Substances 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000010429 evolutionary process Effects 0.000 description 2
- 238000011049 filling Methods 0.000 description 2
- 238000007667 floating Methods 0.000 description 2
- 239000012535 impurity Substances 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000004091 panning Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000013450 outlier detection Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/904—Browsing; Visualisation therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/906—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a display method, a device and a computer readable storage medium for population movement evolution, which comprise the following steps: s1: collecting population movement data; s2: preprocessing the collected population movement data; s3: extracting features from the preprocessed population movement data by using a random forest algorithm, and calculating the keney importance G (k) for the features k; s4: sorting the extracted population migration features according to the importance of the base, selecting the first two features, converting the first two features into an image form, namely mapping the first two features onto pixel values of the image, and generating the image with visual effect by adopting an improved bilinear interpolation algorithm; s5: visual display, namely displaying the evolution process of population movement on display equipment by using the generated image; s6, ending. According to the application, the characteristics are extracted from the preprocessed population movement data by using a random forest algorithm, the image with visual effect is generated by using an improved bilinear interpolation algorithm, and factors such as distance are considered, so that the calculation efficiency is greatly enhanced.
Description
Technical Field
The present invention relates to the field of demographic data statistics, and in particular, to a method and apparatus for displaying population movement evolution, and a computer readable storage medium.
Background
Population migration is an important phenomenon in the current society, and has important significance for city planning and social management. However, the conventional population migration data analysis method has the problems of large information quantity, complex processing and the like, and is difficult to intuitively display the population migration mode and rule. Thus, there is a need for a new method and system that can utilize image processing techniques to convert population migration data into intuitive image form so that a decision maker can better understand and analyze the dynamic changes of population migration.
The population movement evolution display method in the prior art simply performs data presentation according to population movement amount, movement frequency and the like, factors such as population movement distance and the like are not added into the form consideration factors of presentation, the existing population statistics is simply performed according to a single data source, critical extraction is not performed according to important features, data redundancy is poor, computational efficiency is reduced due to excessive feature consideration, accuracy is reduced, and certain robustness is achieved on the statistical data.
Disclosure of Invention
In order to solve the technical problems mentioned in the prior art, the invention provides a display method, a device and a computer readable storage medium for population movement evolution, which are used for extracting features from preprocessed population movement data by utilizing a random forest algorithm and presenting information such as general population flow trend and the like by combining an improved bilinear interpolation algorithm, thereby realizing timely, accurate and dynamic representation of population flow.
The invention discloses a display method of population movement evolution, which comprises the following steps:
s1: collecting population mobile data, storing the collected mobile communication data and social media data into a database, and acquiring call records, position information and check-in records of users;
s2: preprocessing the collected population movement data;
S3: and extracting features from the preprocessed population movement data by using a random forest algorithm, wherein population migration distance features are calculated as follows:
Where d represents the geographic distance, lat 1 and lon 1 represent the latitude and longitude of the first location, lat 2 and lon 2 represent the latitude and longitude of the second location, and R represents the earth radius;
For feature k, the kene importance G (k) can be calculated by the following formula:
G(k)＝∑(p(i)+p(i/k))2
Wherein p (i) represents the frequency of tag i, and p (i/k) represents the frequency of tag i under the condition of feature k;
s4: sorting the extracted population migration features according to the importance of the base, selecting the first two features, converting the first two features into an image form, namely mapping the first two features onto pixel values of the image, and generating the image with visual effect by adopting an improved bilinear interpolation algorithm;
f (x, y) represents the interpolated pixel value, f (0, 0), f (1, 0), f (0, 1), f (1, 1) represent the pixel value of the existing data point, respectively, and α and β represent the offset relative to the existing data point; h is a set adjustment coefficient;
S5: visual display, namely displaying the evolution process of population movement on display equipment by using the generated image; the touch screen interaction realizes zooming, translation and screening operations so as to acquire more detailed information;
S6: and (5) ending.
Preferably, the preprocessing of the collected population movement data includes filtering the population movement data using a time window filter, with the following formula:
F＝(x(n-t)+x(n-t+1)+......x(n))/t
Where x (n) represents the population movement number at the nth time point and t is the window size.
Preferably, the characteristics are extracted from the preprocessed population movement data by using a random forest algorithm, and the characteristics comprise population migration distance, migration intensity, migration frequency information and migration population quantity.
Preferably, the migration frequency information indicates the number or frequency of migration of the population from the start point to the end point, and the migration intensity indication indicates the intensity of flow of the population from the start point to the end point as represented by the population density.
The application also provides a display device for population movement evolution, which comprises:
The data collector collects population movement data from a plurality of data sources, stores the collected mobile communication data and social media data into a database, and acquires call records, position information and check-in records of users;
the preprocessor preprocesses the collected population movement data;
and a feature extractor for extracting features from the preprocessed population movement data by using a random forest algorithm, wherein population migration distance features are calculated as follows:
Where d represents the geographic distance, lat 1 and lon 1 represent the latitude and longitude of the first location, lat 2 and lon 2 represent the latitude and longitude of the second location, and R represents the earth radius;
For feature k, the kene importance G (k) can be calculated by the following formula:
G(k)＝∑(p(i)+p(i/k))2
Wherein p (i) represents the frequency of tag i, and p (i/k) represents the frequency of tag i under the condition of feature k;
the image processor is used for sorting the extracted population migration features according to the importance of the base, selecting the first two features, converting the first two features into an image form, namely mapping the first two features onto pixel values of the image, and generating the image with visual effect by adopting an improved bilinear interpolation algorithm;
f (x, y) represents the interpolated pixel value, f (0, 0), f (1, 0), f (0, 1), f (1, 1) represent the pixel value of the existing data point, respectively, and α and β represent the offset relative to the existing data point; h is a set adjustment coefficient;
The visual display displays the evolution process of population movement on the display device; the touch screen interaction realizes zooming, translation and screening operations so as to acquire more detailed information;
And (5) ending the module.
Preferably, the preprocessing of the collected population movement data includes filtering the population movement data using a time window filter, with the following formula:
F＝(x(n-t)+x(n-t+1)+......x(n))/t
Where x (n) represents the population movement number at the nth time point and t is the window size.
Preferably, the characteristics are extracted from the preprocessed population movement data by using a random forest algorithm, and the characteristics comprise population migration distance, migration intensity, migration frequency information and migration population quantity.
Preferably, the migration frequency information indicates the number or frequency of migration of the population from the start point to the end point, and the migration intensity indication indicates the intensity of flow of the population from the start point to the end point as represented by the population density.
The present invention also provides a computer readable storage medium having stored thereon a population movement evolving display program which when executed by a processor implements the steps of the population movement evolving display method.
The invention provides a display method, a device and a computer readable storage medium for population movement evolution, which can realize the following beneficial technical effects:
1. According to the application, the characteristics are extracted from the preprocessed population movement data by utilizing the random forest algorithm, the improved bilinear interpolation algorithm is adopted to generate the image with the visual effect, and the random forest algorithm and the improved bilinear interpolation algorithm are combined to form a continuous step design, so that the technical scheme is formed, the data judgment accuracy is greatly enhanced, and the data processing efficiency is improved.
2. According to the invention, the geographic distance d is added into the construction process of the image, and is used as one of the influence factors for forming the image, so that the population flow visual effect is greatly enhanced, the pixel value is higher when the population movement distance is longer, and the visual effect is more obvious; meanwhile, the extracted population migration features are ranked according to the importance of the base, the first two features are selected and converted into an image form, namely, the first two features are mapped onto pixel values of the image, and an improved bilinear interpolation algorithm is adopted to generate the image with visual effect;
f (x, y) represents the interpolated pixel value, f (0, 0), f (1, 0), f (0, 1), f (1, 1) represent the pixel value of the existing data point, respectively, and α and β represent the offset relative to the existing data point; h is a set adjustment coefficient; the screening and judging of high-quality data are greatly realized, the data calculation efficiency is enhanced, and the data calculation accuracy is improved.
3. According to the application, the extracted population migration features are ranked according to the importance of the base, the first two features are selected to be converted into the image form, and the feature value with larger influence is selected as the data base for constructing the image, so that the data redundancy is greatly overcome, the calculation efficiency is greatly enhanced, and the rapid real-time display of the population floating display is realized.
4. According to the application, population mobile data are collected from a plurality of data sources, the collected mobile communication data and social media data are stored in the database, call records, position information and check-in records of users are obtained, the richness of the data is fully considered, and the calculation efficiency is greatly enhanced.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram showing the steps of a population movement evolution display method according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1:
In order to solve the above-mentioned problems mentioned in the prior art, as shown in fig. 1: the invention provides a display method of population movement evolution, which comprises the following steps:
s1: collecting population mobile data, storing the collected mobile communication data and social media data into a database, and acquiring call records, position information and check-in records of users;
in some embodiments, the use of conditional random fields (Conditional Random Fields, CRF) is a probabilistic graph model for sequence labeling tasks that can be applied to identify location information in text. The following is a specific example illustration and formulation:
Let us assume that we have a text sequence containing location information that we want to use conditional random fields to identify.
Feature extraction:
First, we need to extract features from the text to be input to the conditional random field model. These features may include parts of speech, contextual information, word boundaries, and the like.
Defining a tag set:
we define a tag set that includes both place (Location) and Non-place (Non-Location) tags.
Defining a characteristic function:
for each combination of observation sequence (input sequence) and tag sequence we define a set of feature functions. The feature function may represent an association of the observation sequence and the tag sequence based on the result of the feature extraction.
Defining a conditional random field model:
the conditional random field model builds a conditional probability distribution between the observation sequence and the tag sequence by parametrically modeling the feature functions.
The formula is:
Given an observation sequence x= { X 1,x2,...,xn }, where X i represents the feature vector of the i-th observation, and a tag sequence y= { Y 1,y2,...,yn }, where Y i represents the i-th tag.
The probability distribution of the conditional random field model is defined as follows:
P(Y|X)＝(1/Z)*exp(∑kλk*∑iTk(yi,yi-1,X,i)+∑kμk*∑iSk(yi,X,i))
Where T k is the transfer characteristic function, S k is the state characteristic function, λ k and μ k are the corresponding weight parameters, and Z is the normalization factor.
The transfer characteristic function Tk measures the relationship between the current tag and the previous tag and the state characteristic function S k measures the relationship between the current tag and the observation.
By training the conditional random field model, optimal weight parameters can be learned, and then a new text sequence is marked, so that place information in the text is identified.
In step 1, various data sources are collected through mobile communication data and social media data, which can be described in detail by the following methods and examples:
And (3) mobile communication data acquisition: the mobile communication data can acquire mobile communication records and location information of the user through cooperation with a telecom operator or a mobile application developer. The following are some specific examples:
co-operating with telecom operators: and in cooperation with a telecom operator, acquiring call records, short message records and base station positioning data of the user. These data may provide the user's communication activity and movement trajectory.
Mobile application data: in cooperation with a mobile application developer, data of a mobile application used by a user is acquired. For example, the user's location information, check-in records, etc. may be obtained by the social media application. Such data may provide social activity and interest preferences of the user.
Social media data collection:
Social media data may be collected through collaboration with a social media platform or using an open API interface. The following are some specific examples:
social media platform collaboration: and in cooperation with the social media platform, acquiring social media data of the user. For example, in cooperation with social media platforms such as WeChat, microblog, QQ and the like, data such as a user's location tag, published dynamics and the like are obtained. Such data may provide social relationships and behavioral characteristics of the user.
Open API interface: and acquiring publicly accessible user data by using an open API interface provided by the social media platform. For example, the text data, the attention relationship, and the like of the user are acquired through the WeChat API. These data may provide the user's opinion and public opinion trends.
S2: preprocessing the collected population movement data; in step 2, the purpose of preprocessing the collected population migration data is to ensure the accuracy and integrity of the data, remove noise and duplicate data for subsequent feature extraction and image processing. The following are examples of some specific pretreatment methods, accompanied by corresponding formulas:
data cleaning:
data cleansing aims at removing abnormal values and invalid data and ensuring the quality of the data. The following are examples of some common data cleansing methods:
Abnormal value detection: abnormal values in the aspects of migration distance, migration time and the like are identified and removed through a statistical method or an outlier detection algorithm.
Missing value processing: for missing population migration data, the corresponding data record may be selected for deletion or for population filling using interpolation methods.
The formula is:
The data cleansing method may be represented by mathematical symbols, for example:
abnormal value detection: if the migration distance d exceeds a certain threshold d_max, d is considered an outlier.
Missing value processing: for missing migration times t, linear interpolation can be used for filling:
t＝(t_prev+t_next)/2
denoising data:
data denoising aims to eliminate random noise in data so as to extract real migration patterns and trends. The following are examples of some common data denoising methods:
Smoothing and filtering: and smoothing the population migration quantity or intensity by using a filtering method such as moving average, weighted average and the like so as to reduce the influence of noise.
Time window filtering: by setting a time window, the average or sum of the migration numbers in the window is calculated to smooth the migration data.
The formula is:
The data denoising method can be represented by mathematical symbols, for example:
smoothing and filtering: the population migration number is smoothed using a moving average filter as follows:
smoothed _value= (x [ n ] +x [ n-1] +, +x [ n-k+1 ])/k, where x [ n ] represents the number of transitions at the nth time point and k is the window size
S3: and extracting features from the preprocessed population movement data by using a random forest algorithm, wherein population migration distance features are calculated as follows:
Where d represents the geographic distance, lat 1 and lon 1 represent the latitude and longitude of the first location, lat 2 and lon 2 represent the latitude and longitude of the second location, and R represents the earth radius;
For feature k, the kene importance G (k) can be calculated by the following formula:
G(k)＝∑(p(i)+p(i/k))2
Wherein p (i) represents the frequency of tag i, and p (i/k) represents the frequency of tag i under the condition of feature k;
in step 3, key features are extracted from the preprocessed population migration data by using a machine learning technology, and a random forest algorithm is one of common methods. The following is a specific calculation process for extracting the characteristics by adopting a random forest algorithm, and a corresponding formula expression is attached:
random forest algorithm introduction:
Random forests are an integrated learning method that performs feature extraction and prediction by constructing multiple decision trees and integrating their results. The random forest has better robustness and accuracy.
The calculation process of random forest feature extraction comprises the following steps:
Assume that our goal is to extract key features from the preprocessed population migration data, including migration distance, migration population number, and migration direction.
A. data preparation:
The marked population migration data is prepared, including migration distance, migration population number and migration direction as features, while target features (key features to be predicted) are used as labels.
B. feature selection:
Suitable features are selected as inputs such as migration distance, migration population number and migration direction.
C. Model training:
and training the model by using a random forest algorithm, and establishing a relation model between the features and the target features. Each decision tree in the random forest is independently trained, and the specific training process is as follows:
A portion of the samples (with the samples put back) are randomly selected from the original data.
A portion of the features is randomly selected from the selected samples (sampled without being put back).
A decision tree is trained using the selected samples and features.
Repeating the steps to construct a plurality of decision trees.
D. Feature importance assessment:
the random forest may provide an assessment of the importance of each feature, the assessment index reflecting the degree of contribution of the feature to the predicted target. Common evaluation criteria include the importance of the base and average reduction in non-purity (MEAN DECREASE Impurity), etc.
The formula is:
Random forest algorithm: random forests integrate the results of multiple decision trees, which can be expressed as:
y=f (x 1, x2,., xn), where y represents the target feature, x1, x2, xn represents the input feature.
Feature importance assessment: random forests can evaluate the importance of features by calculating the average reduced non-purity of individual features. The specific formula is as follows:
average reduced impure = impure of decision tree without the feature-impure of decision tree with the feature
It should be noted that the random forest algorithm can also be used for feature selection, and features with higher importance are selected for subsequent image processing and visual display according to the importance evaluation result of the features.
In the random forest algorithm, f (x 1,x2,...,xn) represents the operation of predicting the input variable x 1,x2,...,xn. The random forest model obtains a final prediction by integrating the prediction results of the plurality of decision tree models.
The specific calculation process is as follows:
Training phase:
For each decision tree:
A new training data set is obtained by randomly performing a put-back sampling (bootstrap sampling) from the original training data set.
A portion of the input features is randomly selected as candidate features for the decision tree.
A decision tree model is trained using the selected training data set and the candidate features.
And obtaining a plurality of decision tree models.
Prediction stage:
For each sample to be predicted:
samples are input into each decision tree for prediction.
And counting the most frequently-occurring category or calculating the average predicted value of the regression tree as the final predicted result of the random forest according to the predicted result of the decision tree.
Specifically, the decision tree prediction process is a process of recursively branching according to the judgment condition of the feature. Each decision tree is trained based on different random sample data and feature subsets and predicted using different features and decision conditions. The random forest can obtain more stable and accurate prediction results by integrating the prediction results of each decision tree.
It should be noted that the specific prediction process and calculation may vary depending on the type of decision tree (classification tree or regression tree). The above description is of the basic principle and calculation process of a random forest algorithm in general. In practical application, the random forest can be further optimized and improved by evaluating the feature importance of the decision tree, adjusting the super parameters of the model and the like.
S4: sorting the extracted population migration features according to the importance of the base, selecting the first two features, converting the first two features into an image form, namely mapping the first two features onto pixel values of the image, and generating the image with visual effect by adopting an improved bilinear interpolation algorithm;
f (x, y) represents the interpolated pixel value, f (0, 0), f (1, 0), f (0, 1), f (1, 1) represent the pixel value of the existing data point, respectively, and α and β represent the offset relative to the existing data point; h is a set adjustment coefficient;
When feature extraction is performed using random forests, the feature's base Importance (Gini Importance) and average reduction non-purity (MEAN DECREASE Impurity) can be calculated to evaluate the Importance of the feature. The following are specific examples of calculations and corresponding formulaic representations:
and (5) calculating the importance of the keni:
The base importance measures the extent to which each feature contributes to the predictive performance of the random forest model. The steps for calculating the importance of the keni are as follows:
for each decision tree, a base Index (Gini Index) is calculated for the features in the decision tree.
The base index of all decision trees is averaged.
The formula is:
For feature k, the kene importance G (k) can be calculated by the following formula:
G(k)＝Σ(p(i)-p(i|k))^2
Where p (i) represents the frequency of tag i and p (i|k) represents the frequency of tag i under the condition of feature k. By calculating the importance of the keni, an importance ranking of the features can be obtained.
Average reduced non-purity calculation:
Average reduction of the degree of uncertainty measures the average degree of each feature used to reduce the degree of uncertainty in the random forest model. The step of calculating the average reduction in non-purity is as follows:
for each decision tree, the reduction in the impure level after splitting in that decision tree using feature k is calculated.
The reduction in the opacity of all decision trees is averaged.
The formula is:
for feature k, average reduced opacity MD (k) can be calculated by the following equation:
MD(k)＝Σ(impurity_before_split-impurity_after_split)
Where, impurity_before_split represents the purity before splitting, impurity_after_split represents the purity after splitting. By calculating the average reduction in the degree of non-purity, the degree of contribution of a feature in the decision tree splitting process can be evaluated.
It should be noted that both the importance of the base and the average reduction in the degree of non-purity are relative indicators that can be used to compare the importance between different features without a standardized range of values. Therefore, in practical applications, the results of the importance assessment should be interpreted and analyzed in connection with specific questions and data sets.
By the above calculation process and formulation, the random forest can be used to calculate the base importance of the feature and average reduction in the non-purity, thereby evaluating the importance of the feature.
In the random forest, when the average reduction of the unrepeatation is calculated, the unrepeatation of the decision tree without the feature and the unrepeatation of the decision tree with the feature need to be calculated respectively. The following are specific examples of calculations and corresponding formulaic representations:
Unreliability of decision tree without this feature:
Examples: assume we have a decision tree for predicting population migration direction. In this decision tree we do not consider the feature "migration distance". We need to calculate the unreliability of decision trees without the "migration distance" feature.
The formula is: assuming that the unrepeace index is I, the unrepeace of the decision tree is indicated. The unrepeatation of the decision tree without the "migration distance" feature may be denoted as i_no_k.
Unreliability of the decision tree with this feature:
Examples: in the same decision tree, we now consider the feature "migration distance". We need to calculate the unreliability of decision trees with "migration distance" features.
The formula is: assume that the decision tree with the "migration distance" feature is not pure i_with_k.
It should be noted that the specific method of calculating the purity depends on the purity index used. In classification problems, commonly used indicators of non-purity include the Gini Index (Gini Index) and entropy (Entropy). The following are examples of calculations and formulations based on the base index in the examples:
And (3) calculating a base index:
for a given node, assume that there are C categories, with the number of samples for each category being C1, C2,..c. The formula for the base index is:
Gini＝1-Σ(ci/n)^2
where n represents the total number of samples.
By calculating the unreliability of the decision tree without this feature (i_no_k) and the unreliability of the decision tree with this feature (i_with_k), an average reduction of unreliability can be calculated. Average reduction of non-purity can be used to evaluate the importance and extent of contribution of features.
The above examples and formulas are based on the calculation of the base index, and the specific calculation method may be different for the case of using other index of non-purity. In practical applications, corresponding calculations may be performed based on the requirements of the problem and the selected index of unrepeatation.
S5: visual display, namely displaying the evolution process of population movement on display equipment by using the generated image; the touch screen interaction realizes zooming, translation and screening operations so as to acquire more detailed information;
S6: and (5) ending.
Preferably, the preprocessing of the collected population movement data includes filtering the population movement data using a time window filter, with the following formula:
F＝(x(n-t)+x(n-t+1)+......x(n))/t
Where x (n) represents the population movement number at the nth time point and t is the window size.
Preferably, the characteristics are extracted from the preprocessed population movement data by using a random forest algorithm, and the characteristics comprise population migration distance, migration intensity, migration frequency information and migration population quantity.
Preferably, the migration frequency information indicates the number or frequency of migration of the population from the start point to the end point, and the migration intensity indication indicates the intensity of flow of the population from the start point to the end point as represented by the population density.
Example 2:
The application also provides a display system for population movement evolution, comprising:
the data acquisition module acquires population mobile data from a plurality of data sources, stores the acquired mobile communication data and social media data into a database, and acquires call records, position information and check-in records of users; in step 1, extracting demographic data features from social media data may employ algorithms of text mining and natural language processing to identify and extract relevant information. The following is a specific description and example:
Text mining algorithm:
Text mining algorithms may be used to extract demographic data features, such as age, gender, geographic location, etc., from social media data. These algorithms are typically based on machine learning and natural language processing techniques, including methods of text classification, named entity recognition, keyword extraction, and the like.
Specific examples are:
assume that we extract age information of a user from social media data. Text classification algorithms, such as naive bayes classifier, may be used to classify the text of the user and determine the age group to which the user belongs.
The formula is:
The naive bayes classifier calculates the probability that the text belongs to a certain specific age group based on bayes theorem.
P (age group |text) =p (text |age group) ×p (age group)/P (text)
Where P (age group |text) represents the probability of an age group given the text, P (text |age group) represents the probability of the text at a particular age group, P (age group) represents the prior probability of the age group, and P (text) represents the probability of the text.
Age information in social media data may be extracted by training a naive bayes classifier and using it to classify new text.
It should be noted that the above examples are directed to specific algorithms and formulas for extracting age characteristics, and that similar text mining and natural language processing techniques may be employed for other demographic data characteristics, such as gender, geographic location, etc., and suitable algorithms and formulas may be selected based on specific questions and data characteristics.
Extracting population migration data features from social media data may employ algorithms of text mining and geographic information processing to identify and extract relevant information. The following is a specific description and example:
Text mining algorithm:
Text mining algorithms may be used to extract relevant information for population migration from social media data, such as location extraction, location association, migration trends, and the like. These algorithms may be based on natural language processing and text analysis techniques, including named entity recognition, location extraction, and semantic analysis.
Specific examples are:
Suppose we extract destination information for population migration from social media data. Named entity recognition algorithms, such as conditional random fields (Conditional Random Fields, CRF) or recurrent neural networks (Recurrent Neural Networks, RNN), may be employed to identify location information in text.
The formula is:
Conditional Random Field (CRF) is a probabilistic graph model that can be used for sequence labeling tasks such as named entity recognition. The goal of CRF is to maximize the conditional probability:
P(Y|X)＝exp(Σw_i*f_i(X,Y))/Z(X)
where Y is the labeling sequence (location information), X is the input text, f_i is the feature function, w_i is the corresponding weight, and Z (X) is the normalization factor.
By training the CRF model and using it to name entity recognition for new text, destination information in social media data can be extracted.
Geographic information processing algorithm:
Geographic information processing algorithms may be used to parse and process the extracted location information to obtain more specific population migration data features such as latitude and longitude coordinates, migration distance, migration trends, and the like. These algorithms may include methods such as geocoding, geodistance calculation, and geographic visualization.
Specific examples are:
suppose we have extracted destination location information in social media data. The location information may be converted to latitude and longitude coordinates using a geocoding service, such as google map API.
The formula is:
Geocoding services typically provide an API interface that can obtain corresponding latitude and longitude coordinates from a place name or address query.
By means of geocoding, destination location information can be converted into computable longitude and latitude coordinates, and characteristics such as population migration distance and population migration trend can be further analyzed.
It should be noted that the above examples are specific algorithms and formulas directed to extracting population migration data features from social media data. Other algorithms for text mining and geographic information processing may also be considered to extract more relevant population migration features, depending on actual needs and data characteristics.
The preprocessing module is used for preprocessing the collected population movement data;
The feature extraction module is used for extracting features from the preprocessed population movement data by using a random forest algorithm, wherein population migration distance features are calculated as follows:
Where d represents the geographic distance, lat 1 and lon 1 represent the latitude and longitude of the first location, lat 2 and lon 2 represent the latitude and longitude of the second location, and R represents the earth radius;
For feature k, the kene importance G (k) can be calculated by the following formula:
G(k)＝∑(p(i)+p(i/k))2
Wherein p (i) represents the frequency of tag i, and p (i/k) represents the frequency of tag i under the condition of feature k;
the image processing module is used for sorting the extracted population migration features according to the importance of the base, selecting the first two features, converting the first two features into an image form, namely mapping the first two features onto pixel values of the image, and generating the image with visual effect by adopting an improved bilinear interpolation algorithm;
f (x, y) represents the interpolated pixel value, f (0, 0), f (1, 0), f (0, 1), f (1, 1) represent the pixel value of the existing data point, respectively, and α and β represent the offset relative to the existing data point; h is a set adjustment coefficient;
The extracted population migration features are converted into image forms in order to generate images with visual effects through an image processing algorithm so as to more intuitively show the population migration modes and trends. The following is a specific example illustration:
Converting into an image form:
Suppose we extract two features from population migration data: migration distance and migration population number. We can map these two features onto the pixel values of the image. One common approach is to map migration distances to the horizontal axis of the image and migration population numbers to the vertical axis of the image. Thus, the brightness or color of each pixel point may represent the values of the migration distance and the number of migration population at the corresponding location.
Generating an image processing algorithm:
one common image processing algorithm is an interpolation algorithm that fills the entire image by interpolating pixels between existing data points. This may make the image smoother and provide more detail. Common interpolation algorithms include nearest neighbor interpolation, bilinear interpolation, bicubic interpolation, and the like.
Specific examples are:
Assume that we extract the following features from population migration data:
migration distance: [100,200,150,300]
Migration population number: [500,1000,800,1200]
We can map migration distances to the horizontal axis of the image and migration population numbers to the vertical axis of the image. Assuming that the width and the height of the image are 400 pixels, respectively, the lateral and longitudinal corresponding distances of each pixel are 1 pixel.
The image may then be processed using an interpolation algorithm to generate an image with a visual effect. Taking bilinear interpolation as an example, the interpolation algorithm will calculate new pixel values based on the values of the existing data points.
The visual display module displays the evolution process of population movement on the display equipment by the generated image; the touch screen interaction realizes zooming, translation and screening operations so as to acquire more detailed information;
In step 5, algorithms such as scaling, translation, and screening operations may be employed in order for the decision maker to intuitively observe and analyze the evolution of population movement and to interoperate to obtain more detailed information. The following is a specific example illustration:
scaling algorithm:
The scaling algorithm is used to adjust the display scale of the image so that a decision maker can observe different levels of detail. One common scaling algorithm is bilinear interpolation, which can generate new pixel values by calculating a weighted average of neighboring pixels.
The formula is:
The bilinear interpolation formula is as follows:
f(x,y)＝(1-α)(1-β)f(0,0)+α(1-β)f(1,0)+(1-α)βf(0,1)+αβf(1,1)
Where f (x, y) represents the interpolated pixel value, f (0, 0), f (1, 0), f (0, 1) and f (1, 1) represent the pixel value on the original image, respectively, and α and β represent the offset relative to the original pixel.
Translation algorithm:
The panning algorithm is used to perform panning operations on the image so that a decision maker can move the image and view the region of interest. The translation algorithm simply translates the pixels on the image along a specified direction.
The formula is:
for the translation operation, this can be achieved by adjusting the coordinates of the pixels. For example, for a shift of tx pixels to the right, the new coordinates of the pixel are (x+tx, y).
Screening algorithm:
the screening algorithm is used to screen out the portion of interest from the image according to specific conditions. This can be achieved by setting a threshold value or using color characteristics of the pixels, etc.
The formula is:
the specific formula of the screening algorithm depends on the conditions and method used. For example, if we want to filter out portions with pixel values greater than the threshold, the following formula can be used:
filtered_image(x,y)＝
f(x,y),if f(x,y)>threshold
0,otherwise
where f (x, y) represents a pixel value on the original image, and threshold represents a set threshold value.
Through the above interactive algorithm, the decision maker can freely zoom, pan and filter the image to obtain more detailed information. This can help the decision maker to gain insight into the evolutionary process of population movement and make more accurate decisions. Appropriate algorithms and formulas need to be selected according to specific application scenarios and requirements.
And (5) ending the module.
Preferably, the preprocessing of the collected population movement data includes filtering the population movement data using a time window filter, with the following formula:
F＝(x(n-t)+x(n-t+1)+......x(n))/t
Where x (n) represents the population movement number at the nth time point and t is the window size.
Preferably, the characteristics are extracted from the preprocessed population movement data by using a random forest algorithm, and the characteristics comprise population migration distance, migration intensity, migration frequency information and migration population quantity.
Preferably, the migration frequency information indicates the number or frequency of migration of the population from the start point to the end point, and the migration intensity indication indicates the intensity of flow of the population from the start point to the end point as represented by the population density.
In the random forest algorithm, f (x 1,x2,...,xn) represents an operation of integrated prediction by a plurality of decision trees. The specific calculation process is as follows:
For each decision tree:
The input variable x 1,x2,...,xn is used as an input to the decision tree.
The decision tree makes decisions and branches based on the input variables and the tree structure until leaf nodes are reached.
The leaf node contains a predictor or class label.
The decision tree branches according to the characteristic value of the input variable through judging conditions, and predicts according to the predicted value or class label of the leaf node.
In the random forest, f (x 1,x1,...,xn) represents an operation by integrating the prediction results of a plurality of decision trees. Common ways of integration include taking an average of the predicted results (for regression problems) or voting (for classification problems).
The concrete steps are as follows:
For regression problems:
f(x1,x2,...,xn)＝(1/M)*Σifi(x1,x2,...,xn)
where M represents the number of decision trees in the random forest, and f i(x1,x1,...,xn) represents the predicted outcome of the ith decision tree.
For classification problems:
f(x1,x2,...,xn)＝argmax(ΣiCount(y＝c|fi(x1,x2,...,xn)))
Wherein argmax represents the category c that maximizes the value in brackets, and Count (y=c|f i(x1,x2,...,xn)) represents the number of occurrences of category c in the prediction result of the i-th decision tree.
It should be noted that the predicted outcome of each decision tree in the random forest may be a continuous value (regression problem) or a discrete class label (classification problem), and the final integrated predicted outcome f (x 1,x2,...,xn) is derived from the statistical information of the predicted outcome of the decision tree. The specific prediction mode may vary from problem to problem.
The average value ,f(x1,x2,...,xn)＝(1/M)*Σifi(x1,x2,...,xn) in the above scheme represents the average of the predicted results of multiple decision trees in the random forest. The prediction result may be a continuous value in the regression problem or a discrete class label in the classification problem.
Specifically, after preprocessing and feature extraction of the mobile communication data and the social media data, a set of feature vectors (x 1,x1,...,xn) is obtained. These feature vectors contain important information describing population migration, such as the origin of population flow, destination, time of migration, distance of migration, etc. These feature vectors are passed as inputs to a random forest model for prediction.
Each decision tree f i(x1,x2,...,xn in the random forest) is predicted independently based on the input feature vector. For regression problems, the prediction result of the decision tree may be a continuous value, such as predicting the number of population shifts. For classification problems, the predictive outcome of the decision tree may be a discrete category label, such as predicting the type of population migration (e.g., job migration, educational migration, etc.).
The overall prediction result can be obtained by averaging the prediction result of each decision tree in the random forest. This integration helps to reduce the prediction error of a single decision tree and improves overall prediction accuracy. The prediction results may help decision makers understand trends and laws of population migration, thereby supporting decision making and planning.
In summary, the mean f (x 1,x2,...,xn) can be regarded as a comprehensive prediction result of the feature information extracted by the random forest model on the mobile communication data and the social media data. By modeling and predicting the feature information, insight into population migration can be gained, providing decision support for decision makers.
In the above scheme, other calculations may be performed in addition to the mean calculation to obtain more information about the predicted result. The following are some possible examples of calculations:
variance calculation:
The variance can measure the discrete degree of the prediction results of a plurality of decision trees in the random forest, and reflects the stability and consistency of the prediction results. The variance may be obtained by calculating an average of the squared differences of each sample over a plurality of decision tree predictors.
The formula is:
Var(y)＝(1/M)*Σi(fi(x1,x2,...,xn)-f(x1,x2,...,xn))2
Probability calculation:
for classification problems, the predictive probability for each category may be calculated to measure the likelihood of each category. The probability can be obtained by counting the number of occurrences of each category in the random forest and dividing by the total number of decision trees.
The formula is:
P(y＝c)＝Count(y＝c)/M
These calculations may provide a more comprehensive analysis of the predicted results, helping to further understand and explain the predicted effects of the random forest model. Variance calculations may measure the stability of the predicted outcome, while probability calculations may provide confidence information for each category. These additional calculations may provide the decision maker with more insight and decision support regarding population migration.
In a random forest decision tree, the mean calculation, variance calculation and probability calculation represent in particular different statistical analyses of the predicted outcome. They can provide information about different aspects of the predicted outcome, helping us to better understand and interpret the behavior of the model.
And (3) calculating a mean value:
The mean calculation represents an average of the predictions for a plurality of decision trees. The method is used for obtaining the overall prediction result and can be applied to continuous value prediction in regression problems or category prediction in classification problems. The purpose of the mean value calculation is to reduce the prediction error of a single decision tree in a random forest and improve the overall prediction accuracy.
Variance calculation:
variance calculation is used to measure the degree of dispersion of the predicted results of multiple decision trees in a random forest. It may reflect the stability and consistency of the predicted outcome. The variance can be obtained by calculating the average of the square differences of each sample over multiple decision tree predictors. A smaller variance indicates that the prediction results of the decision tree are more consistent, and a larger variance indicates that the prediction results have a larger variance.
Probability calculation:
The probability calculation is used to obtain a predicted probability for each category in the classification problem. The predictive probability for each category can be obtained by counting the number of occurrences of each category in the random forest and dividing by the total number of decision trees. Probability calculations can help us understand the size of the likelihood of each category to better understand and interpret the prediction.
The relationship with the evolution process of the generated image in the subsequent step to show population movement on the display device is that the statistical analysis results (mean, variance, probability) can be used as the basis or reference for generating the image. Depending on the statistics, different image processing methods may be selected to demonstrate the evolutionary process of population movement. For example, the mean may be used to generate a thermodynamic diagram showing changes in population density, the variance may be used to show uncertainty in population flow, and the probability may be used to generate a classification map showing the distribution of different categories. These statistical analysis results provide basic data and guidance to help generate images with visual effects so that a decision maker can intuitively observe and analyze the evolving process of population movement and perform further operations and analysis in an interactive manner.
The invention provides a display system for population movement evolution, which has the following beneficial technical effects:
1. According to the application, the characteristics are extracted from the preprocessed population movement data by utilizing the random forest algorithm, the improved bilinear interpolation algorithm is adopted to generate the image with the visual effect, and the random forest algorithm and the improved bilinear interpolation algorithm are combined to form a continuous step design, so that the technical scheme is formed, the data judgment accuracy is greatly enhanced, and the data processing efficiency is improved.
2. According to the invention, the geographic distance d is added into the construction process of the image, and is used as one of the influence factors for forming the image, so that the population flow visual effect is greatly enhanced, the pixel value is higher when the population movement distance is longer, and the visual effect is more obvious; meanwhile, the extracted population migration features are ranked according to the importance of the base, the first two features are selected and converted into an image form, namely, the first two features are mapped onto pixel values of the image, and an improved bilinear interpolation algorithm is adopted to generate the image with visual effect;
f (x, y) represents the interpolated pixel value, f (0, 0), f (1, 0), f (0, 1), f (1, 1) represent the pixel value of the existing data point, respectively, and α and β represent the offset relative to the existing data point; h is a set adjustment coefficient; the screening and judging of high-quality data are greatly realized, the data calculation efficiency is enhanced, and the data calculation accuracy is improved.
3. According to the application, the extracted population migration features are ranked according to the importance of the base, the first two features are selected to be converted into the image form, and the feature value with larger influence is selected as the data base for constructing the image, so that the data redundancy is greatly overcome, the calculation efficiency is greatly enhanced, and the rapid real-time display of the population floating display is realized.
4. According to the application, population mobile data are collected from a plurality of data sources, the collected mobile communication data and social media data are stored in the database, call records, position information and check-in records of users are obtained, the richness of the data is fully considered, and the calculation efficiency is greatly enhanced.
The foregoing has outlined a detailed description of a method for obtaining demographic data, wherein specific examples are provided herein to illustrate the principles and embodiments of the present invention, the above examples being provided solely to assist in understanding the core concept of the present invention; also, as will be apparent to those skilled in the art in light of the present teachings, the present disclosure should not be limited to the specific embodiments and applications described herein.
Claims (9)
1. A method for displaying the mobile evolution of a population, comprising the steps of:
s1: collecting population mobile data, storing the collected mobile communication data and social media data into a database, and acquiring call records, position information and check-in records of users;
s2: preprocessing the collected population movement data;
S3: and extracting features from the preprocessed population movement data by using a random forest algorithm, wherein population migration distance features are calculated as follows:
)
Wherein, Representing geographical distance,/>And/>Representing latitude and longitude of the first place,/>And/>A latitude and longitude representing a second location, R representing an earth radius;
for feature k, the kene importance G (k) is calculated by the following formula:
Wherein, Representing the frequency of tag i,/>Representing the frequency of tag i under the condition of feature k;
s4: sorting the extracted population migration features according to the importance of the base, selecting the first two features, converting the first two features into an image form, namely mapping the first two features onto pixel values of the image, and generating the image with visual effect by adopting an improved bilinear interpolation algorithm;
Representing interpolated pixel values,/> 、/>、/>、/>Respectively representing pixel values of existing data points, and alpha and beta representing offsets relative to the existing data points; h is a set adjustment coefficient;
S5: visual display, namely displaying the evolution process of population movement on display equipment by using the generated image; the touch screen interaction realizes zooming, translation and screening operations so as to acquire more detailed information;
S6: and (5) ending.
2. The method of claim 1, wherein preprocessing the collected population movement data comprises filtering the population movement data using a time window filter, the formula being:
Wherein, Represents the population movement number at the nth time point, and t is the window size.
3. The method of claim 2, wherein the feature is extracted from the preprocessed population movement data using a random forest algorithm, and the feature includes population migration distance, migration intensity, migration frequency information, and migration population number.
4. A method of displaying a population movement evolution as in claim 3, wherein the migration frequency information indicates a number or frequency of population migration from a start point to an end point, and wherein the migration intensity indicates a population flow intensity from the start point to the end point as represented by a population density.
5. A display device for the mobile evolution of a population, comprising:
The data collector collects population movement data from a plurality of data sources, stores the collected mobile communication data and social media data into a database, and acquires call records, position information and check-in records of users;
the preprocessor preprocesses the collected population movement data;
and a feature extractor for extracting features from the preprocessed population movement data by using a random forest algorithm, wherein population migration distance features are calculated as follows:
)
Wherein, Representing geographical distance,/>And/>Representing latitude and longitude of the first place,/>And/>A latitude and longitude representing a second location, R representing an earth radius;
for feature k, the kene importance G (k) is calculated by the following formula:
Wherein, Representing the frequency of tag i,/>Representing the frequency of tag i under the condition of feature k;
the image processor is used for sorting the extracted population migration features according to the importance of the base, selecting the first two features, converting the first two features into an image form, namely mapping the first two features onto pixel values of the image, and generating the image with visual effect by adopting an improved bilinear interpolation algorithm;
Representing interpolated pixel values,/> 、/>、/>、/>Respectively representing pixel values of existing data points, and alpha and beta representing offsets relative to the existing data points; h is a set adjustment coefficient;
The visual display module displays the evolution process of population movement on the display equipment by the generated image; the touch screen interaction realizes zooming, translation and screening operations so as to acquire more detailed information;
And (5) ending the module.
6. The population movement evolving display device of claim 5, wherein preprocessing the collected population movement data comprises filtering the population movement data using time window filtering, formulated as follows:
Wherein, Represents the population movement number at the nth time point, and t is the window size.
7. The evolving display device of population movement of claim 5, wherein the features extracted from the preprocessed population movement data using random forest algorithm include population migration distance, migration intensity, migration frequency information, migration population number.
8. The display of population movement evolution of claim 7, wherein the migration frequency information indicates a number or frequency of population migration from a start point to an end point, and wherein the migration intensity indicates a population flow intensity from a start point to an end point as represented by a population density.
9. A computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, which when executed by a processor, implements the steps of the method of displaying the mobile evolution of the population according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310730446.9A CN117112859B (en) | 2023-06-16 | 2023-06-16 | Display method, device and computer readable storage medium for population movement evolution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310730446.9A CN117112859B (en) | 2023-06-16 | 2023-06-16 | Display method, device and computer readable storage medium for population movement evolution |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117112859A CN117112859A (en) | 2023-11-24 |
| CN117112859B true CN117112859B (en) | 2024-05-14 |
Family
ID=88806258
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310730446.9A Active CN117112859B (en) | 2023-06-16 | 2023-06-16 | Display method, device and computer readable storage medium for population movement evolution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117112859B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096631A (en) * | 2016-06-02 | 2016-11-09 | 上海世脉信息科技有限公司 | A kind of recurrent population's Classification and Identification based on the big data of mobile phone analyze method |
| WO2020233259A1 (en) * | 2019-07-12 | 2020-11-26 | 之江实验室 | Multi-center mode random forest algorithm-based feature importance sorting system |
| CN113743453A (en) * | 2021-07-21 | 2021-12-03 | 东北大学 | Population quantity prediction method based on random forest |
-
2023
- 2023-06-16 CN CN202310730446.9A patent/CN117112859B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106096631A (en) * | 2016-06-02 | 2016-11-09 | 上海世脉信息科技有限公司 | A kind of recurrent population's Classification and Identification based on the big data of mobile phone analyze method |
| WO2020233259A1 (en) * | 2019-07-12 | 2020-11-26 | 之江实验室 | Multi-center mode random forest algorithm-based feature importance sorting system |
| CN113743453A (en) * | 2021-07-21 | 2021-12-03 | 东北大学 | Population quantity prediction method based on random forest |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117112859A (en) | 2023-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113302634B (en) | System, medium, and method for learning and predicting key phrases and generating predictions | |
| US11521221B2 (en) | Predictive modeling with entity representations computed from neural network models simultaneously trained on multiple tasks | |
| WO2022217839A1 (en) | Air quality prediction method based on deep spatiotemporal similarity | |
| CN111061946A (en) | Scenario content recommendation method and device, electronic equipment and storage medium | |
| CN112819099B (en) | Training method, data processing method, device, medium and equipment for network model | |
| CN111695046B (en) | User portrait inference method and device based on spatio-temporal mobile data representation learning | |
| CN112905897B (en) | Similar user determination method, vector conversion model, device, medium and equipment | |
| CN116611546B (en) | Knowledge-graph-based landslide prediction method and system for target research area | |
| CN117237574B (en) | Task-driven geographical digital twin scene enhancement visualization method and system | |
| WO2020174233A1 (en) | Machine-learned model selection network planning | |
| CN115359873B (en) | Control method for operation quality | |
| CN112036659A (en) | Social network media information popularity prediction method based on combination strategy | |
| CN115422441A (en) | Continuous interest point recommendation method based on social space-time information and user preference | |
| KR20210083089A (en) | Method and system for calculating tourist index | |
| Turgut et al. | A framework proposal for machine learning-driven agent-based models through a case study analysis | |
| CN111161238A (en) | Image quality evaluation method and device, electronic device, and storage medium | |
| CN117746546A (en) | Service business handling method and system based on number calling device | |
| CN113590971A (en) | Interest point recommendation method and system based on brain-like space-time perception characterization | |
| CN117112859B (en) | Display method, device and computer readable storage medium for population movement evolution | |
| CN117540336A (en) | Time sequence prediction method and device and electronic equipment | |
| KR102014288B1 (en) | Development pressure prediction method based on artificial intelligence using drone | |
| US20050256813A1 (en) | Method and system for data understanding using sociomapping | |
| CN111027771A (en) | Scenic spot passenger flow volume estimation method, system and device and storable medium | |
| CN116910341A (en) | Label prediction method and device and electronic equipment | |
| Adiga et al. | Enhancing COVID-19 ensemble forecasting model performance using auxiliary data sources |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |