CN112862195B - SFT-ALS-based time series vermicelli fluctuation prediction method - Google Patents
SFT-ALS-based time series vermicelli fluctuation prediction method Download PDFInfo
- Publication number
- CN112862195B CN112862195B CN202110188635.9A CN202110188635A CN112862195B CN 112862195 B CN112862195 B CN 112862195B CN 202110188635 A CN202110188635 A CN 202110188635A CN 112862195 B CN112862195 B CN 112862195B
- Authority
- CN
- China
- Prior art keywords
- vermicelli
- matrix
- als
- video
- feature
- 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
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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Biophysics (AREA)
- Algebra (AREA)
- Computing Systems (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Fuzzy Systems (AREA)
- Evolutionary Computation (AREA)
- Development Economics (AREA)
- Artificial Intelligence (AREA)
- Game Theory and Decision Science (AREA)
- Genetics & Genomics (AREA)
- Physiology (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
Abstract
The method comprises the following steps of 1, obtaining user basic data; step 2, extracting basic data features, including: daily video update amount, vermicelli amount, video praise amount, video collection amount, video play amount, average growth rate of vermicelli and SFT characteristics of vermicelli time sequence signals; step 3, carrying out normalization processing on the characteristics; step 4, the feature sets after feature processing are formed into a mixed matrix, the mixed matrix is sent into an ALS model for matrix decomposition, and the mixed matrix is reconstructed; step 5, optimizing by using a genetic algorithm to obtain the dimension k of the optimal low-dimensional matrix; and step 6, embedding the vermicelli expansion model obtained by training into a platform system, and applying the vermicelli expansion model practically. The invention provides a time sequence vermicelli expansion prediction method based on SFT-ALS. The vermicelli expansion amplitude is predicted and analyzed by extracting various time sequence characteristics related to the vermicelli expansion amplitude and combining with an ALS model, so that the rules and the characteristics of each wizard worker in the process of the wizard development are revealed.
Description
Technical Field
The invention relates to the field of vermicelli expansion prediction, in particular to a time sequence vermicelli expansion prediction method based on SFT-ALS.
Background
Big data technology originated in the 90 s of the 20 th century, theory published two papers, *** file system and cluster-based simple data processing, mapReduce, with *** publication matured in 2008, after which this technology was widely applied in various fields of business, science, medical science, government, education, economy, traffic, logistics and society. With the continuous maturation of data theory and continuous progress of data acquisition, data research tools are continuously perfected, and data analysis work is changed from a large company to a small team.
Thus we want to put the strength of science and technology into the culture creation through this project. At present, culture creators on various flow platforms (such as sound shaking, fast hands and beeping knotweed) have to select materials with flow when facing a large number of audiences and other competitive wizards, and build personalized contents in various presentation modes, artistic expression modes and the like so as to improve the identification degree of individual ips. We can find out that a good work of the literary composition is heavy in work load and huge in difficulty. First, the literature workers have to cut viewer preferences daily and also find and use the quality of the best-effort hot material in order to select an appropriate material. Secondly, in order to increase the content acceptance and satisfaction of the audience to the originality workers, the originality workers have to spend a great deal of time "packaging" the content, making it more interesting and easier for the audience to accept. Moreover, the current flow platform is provided with a set of content pushing mode, and the content which is easy to be accepted by audiences is pushed preferentially by the mode, so that the literature creator faces the dilemma of low viscosity of target audiences and unstable content income. Especially, the problems that content similarity of the literary works is high, style identification is low, audience groups are extremely unstable, personal IP is difficult to discover and the like are necessarily caused along with the large amount of rushes of individual literary workers.
The system provides more scientific and practical analysis for the literary composition workers through customized analysis of personal IP, and provides scientific basis for customizing the creation route for the literary composition workers.
Disclosure of Invention
In order to solve the problems, the invention provides a time sequence vermicelli fluctuation prediction method based on SFT-ALS on the basis of data acquisition of a system platform. The vermicelli expansion amplitude is predicted and analyzed by extracting various time sequence characteristics related to the vermicelli expansion amplitude and combining with an ALS model, so that the rules and the characteristics of each wizard worker in the process of the wizard development are revealed. The invention provides a time sequence vermicelli expansion prediction method based on SFT-ALS, which comprises the following specific steps:
step 3, data preprocessing: in order to reduce the model training time, carrying out normalization processing on the collected user basic data;
step 4, the feature sets after feature processing are formed into a mixed matrix, the mixed matrix is sent into an ALS model for matrix decomposition, and the mixed matrix is reconstructed;
step 5, optimizing training models of different data sources by using a genetic algorithm to obtain the dimension k of the optimal low-dimensional matrix;
and step 6, embedding the vermicelli expansion model obtained by training into a platform system, and applying the vermicelli expansion model practically.
Further, the process of data feature extraction in step 2 may be expressed as:
the invention takes daily video update quantity, vermicelli quantity, video praise quantity, video collection quantity and video play quantity as the characteristics of the model, and extracts the average growth rate of vermicelli, and the formula is as follows:
wherein B is the current vermicelli number, A is the first-day vermicelli number, and m is the annual average growth rate.
Further, the process of SFT extraction in step 2 can be expressed as:
the invention extracts SFT characteristics of vermicelli simultaneously, takes vermicelli time sequence signal as e (t) and slow characteristic variable as s i (t), i is the signal dimension, and the slow features are calculatedThe optimization problem of the method is converted into:
in the method, in the process of the invention,<…>which means that the time is averaged over the time,
is the first derivative of slow features, and is obtained by linear conversion of feature variables:
in the formula g i (e) Is a mapping function in a slow feature algorithm, w i Is a load matrix, which is obtained after operations such as whitening treatment, singular value decomposition and the like on vermicelli time sequence signals, and slow characteristics s of the vermicelli time sequence signals are obtained i 。
Further, the process of data preprocessing in step 3 can be expressed as:
in order to reduce training time of the model, the video update amount, the fan amount, the video praise amount, the video collection amount, the video play amount, the fan average growth rate and the fan SFT characteristics of the user extracted in the step 2 are respectively normalized, the normalized characteristic value interval is [ -1,1], and the normalization formula is as follows:
wherein, x' is the feature matrix after normalization processing, x is the feature matrix extracted in the step 2, x max And x min Maximum and minimum matrices of features, respectively.
Further, the process of reconstructing the mixing matrix by the ALS model in step 4 may be expressed as:
combining the normalized feature matrix and daily vermicelli increment into a mixed matrix R m*n (n is the number of features plus 1, m is the number of days of use of the user data);
step 4.1: the mixing matrix is approximately set to
Wherein X is m*k And Y T n*k Is R m*n The low-dimensional matrix obtained by decomposition, wherein the parameter k is the dimension of the low-dimensional matrix;
step 4.2: to make the two sides of the equation as equal as possible, a square error loss function of the alternating least squares method is constructed:
wherein R is ui 、X u 、Y i R is respectively m*n 、X m*k 、Y n*k L (X, Y) is a square error loss function;
step 4.3, at this point the matrix decomposition problem can be translated into a solution with minimum square error loss function:
the key to the ALS algorithm is to find the optimal X m*k And Y n*k Minimize L (X, Y) due to X m*k And Y n*k Is unknown, so the solution problem of the formula is not convex, the thought of the ALS algorithm is to fix one matrix and then solve the other matrix, and the specific algorithm is as follows:
step 4.3.1 firstly taking a random value to fix X u ；
Step 4.3.2 solving for Y for L (X, Y) i And let the bias guide be 0, can solve Y i
Y i ＝(X T X+λI) -1 X T R i (8)
Step 4.3.3 solving the aboveY of solution i Fixed, X can be found by the same principle u
X u ＝(Y T Y+λI) -1 Y T R u (9)
Step 4.3.4 continuously repeats steps 4.3.2 and 4.3.3 until L (X, Y) reaches a target value or a maximum number of iterations is reached;
initially X u And Y i Is a random matrix, and the alternate least square method modifies X by continuous alternate iteration u And Y i Thereby obtaining the final X u And Y i Is a value of (2); solving for X u And Y i And reconstructing the mixed matrix by the formula 5, wherein the reconstructed mixed matrix contains the future expansion quantity of the vermicelli to be solved.
Further, the genetic algorithm optimizing process in step 5 may be expressed as:
k＝GA(x′，ALS) (10)
where x' is the feature matrix of the dataset, ALS represents the ALS model trained by the dataset, and GA (·) is a genetic algorithm function.
The SFT-ALS-based time sequence vermicelli expansion prediction method has the beneficial effects that: the invention has the technical effects that:
1. the invention predicts the expansion condition of the user vermicelli by various time sequence characteristics, thereby improving the accuracy of the model;
2. the invention adopts the genetic algorithm to obtain the optimal parameters of the model, thereby increasing the accuracy of the model;
3. the invention predicts and analyzes the expansion of the vermicelli, and reveals the rules and characteristics of each culture creator in the culture development process.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a graph of the growth analysis of the system fan of the present invention.
Detailed Description
The invention is described in further detail below with reference to the attached drawings and detailed description:
the invention provides a time sequence vermicelli expansion prediction method based on SFT-ALS, which aims to predict and analyze vermicelli expansion by a platform system and reveal rules and characteristics of each culture creator in the culture creating development process. FIG. 1 is a flow chart of the present invention. The steps of the present invention will be described in detail with reference to the flow charts.
the process of data feature extraction in step 2 can be expressed as:
the invention takes daily video update quantity, vermicelli quantity, video praise quantity, video collection quantity and video play quantity as the characteristics of the model, and extracts the average growth rate of vermicelli, and the formula is as follows:
wherein B is the current vermicelli number, A is the first-day vermicelli number, and m is the annual average growth rate.
The process of SFT extraction in step 2 can be expressed as:
the invention extracts SFT characteristics of vermicelli simultaneously, takes vermicelli time sequence signal as e (t) and slow characteristic variable as s i (t), i is the signal dimension, converting the optimization problem of the slow feature algorithm into:
in the method, in the process of the invention,<…>which means that the time is averaged over the time,
is the first derivative of slow features, and is obtained by linear conversion of feature variables:
in the formula g i (e) Is a mapping function in a slow feature algorithm, w i Is a load matrix, which is obtained after operations such as whitening treatment, singular value decomposition and the like on vermicelli time sequence signals, and slow characteristics s of the vermicelli time sequence signals are obtained i 。
Step 3, data preprocessing: in order to reduce the model training time, carrying out normalization processing on the collected user basic data;
the process of data preprocessing in step 3 can be expressed as:
in order to reduce training time of the model, the video update amount, the fan amount, the video praise amount, the video collection amount, the video play amount, the fan average growth rate and the fan SFT characteristics of the user extracted in the step 2 are respectively normalized, the normalized characteristic value interval is [ -1,1], and the normalization formula is as follows:
wherein, x' is the feature matrix after normalization processing, x is the feature matrix extracted in the step 2, x max And x min Maximum and minimum matrices of features, respectively.
Step 4, the feature sets after feature processing are formed into a mixed matrix, the mixed matrix is sent into an ALS model for matrix decomposition, and the mixed matrix is reconstructed;
the process of reconstructing the mixing matrix by the ALS model in step 4 can be expressed as:
combining the normalized feature matrix and daily vermicelli increment into a mixed matrix R m*n (n is the number of features plus 1, m is the number of days of use of the user data);
step 4.1: the mixing matrix is approximately set to
Wherein X is m*k And Y T n*k Is R m*n The low-dimensional matrix obtained by decomposition, wherein the parameter k is the dimension of the low-dimensional matrix;
step 4.2: to make the two sides of the equation as equal as possible, a square error loss function of the alternating least squares method is constructed:
wherein R is ui 、X u 、Y i R is respectively m*n 、X m*k 、Y n*k L (X, Y) is a square error loss function;
step 4.3, at this point the matrix decomposition problem can be translated into a solution with minimum square error loss function:
the key to the ALS algorithm is to find the optimal X m*k And Y n*k Minimize L (X, Y) due to X m*k And Y n*k Is unknown, so the solution problem of the formula is not convex, the thought of the ALS algorithm is to fix one matrix and then solve the other matrix, and the specific algorithm is as follows:
step 4.3.1 firstly taking a random value to fix X u ；
Step 4.3.2 solving for Y for L (X, Y) i And let the bias guide be 0, can solve Y i
Y i ＝(X T X+λI) -1 X T R i (8)
Step 4.3.3 solving the above-mentioned Y i Fixed, X can be found by the same principle u
X u ＝(Y T Y+λI) -1 Y T R u (9)
Step 4.3.4 continuously repeats steps 4.3.2 and 4.3.3 until L (X, Y) reaches a target value or a maximum number of iterations is reached;
initially X u And Y i Is a random matrix, and the alternate least square method modifies X by continuous alternate iteration u And Y i Thereby obtaining the final X u And Y i Is a value of (2); solving for X u And Y i And reconstructing the mixed matrix by the formula 5, wherein the reconstructed mixed matrix contains the future expansion quantity of the vermicelli to be solved, and displaying the calculation result by the system through an interface chart, as shown in figure 2.
Step 5, optimizing training models of different data sources by using a genetic algorithm to obtain the dimension k of the optimal low-dimensional matrix;
the genetic algorithm optimizing process in step 5 can be expressed as:
k＝GA(x′，ALS) (10)
where x' is the feature matrix of the dataset, ALS represents the ALS model trained by the dataset, and GA (·) is a genetic algorithm function.
And step 6, embedding the vermicelli expansion model obtained by training into a platform system, and applying the vermicelli expansion model practically.
The above description is only of the preferred embodiment of the present invention, and is not intended to limit the present invention in any other way, but is intended to cover any modifications or equivalent variations according to the technical spirit of the present invention, which fall within the scope of the present invention as defined by the appended claims.
Claims (2)
1. The SFT-ALS-based time sequence vermicelli expansion prediction method comprises the following specific steps:
step 1, obtaining user basic data: after authorized permission, the system platform collects the basic information of the user, and the number of videos, the expanding data of the vermicelli, the video praise, the video collection and the video play quantity released by the user every day in the past;
step 2, extracting data characteristics: obtaining the essential characteristics of slow data change by using SFT, and extracting the average growth rate characteristics of the data;
the process of data feature extraction in step 2 is represented as: except for daily video update amount, fan amount, video praise amount, video collection amount and video play amount
As the characteristics of the model, the average growth rate of the vermicelli is extracted, and the formula is as follows:
wherein B is the current vermicelli number, A is the first-day vermicelli number, and m is the annual average growth rate;
the process of SFT extraction in step 2 is expressed as:
simultaneously extracting SFT characteristics of vermicelli, setting vermicelli time sequence signal as e (t) and slow characteristic variable as s i (t), i is the signal dimension, converting the optimization problem of the slow feature algorithm into:
where <.> represents averaging over time, s is the slow feature first derivative, and linear conversion of the feature variable is available:
in the formula g i (e) Is a mapping function in a slow feature algorithm, w i Is a load matrix, and is processed by vermicelli
Whitening treatment of time sequence signals, obtaining a load matrix after singular value decomposition operation, and obtaining slow characteristics s of vermicelli time sequence signals i ；
Step 3, data preprocessing: in order to reduce the model training time, carrying out normalization processing on the collected user basic data;
the process of data preprocessing in step 3 is expressed as:
in order to reduce training time of the model, the video update amount, the fan amount, the video praise amount, the video collection amount, the video play amount, the fan average growth rate and the fan SFT characteristics of the user extracted in the step 2 are respectively normalized, the normalized characteristic value interval is [ -1,1], and the normalization formula is as follows:
wherein, x' is the feature matrix after normalization processing, x is the feature matrix extracted in the step 2, x max And x min A maximum value matrix and a minimum value matrix of the features respectively;
step 4, the feature sets after feature processing are formed into a mixed matrix, the mixed matrix is sent into an ALS model for matrix decomposition, and the mixed matrix is reconstructed;
the process of reconstructing the mixing matrix by the ALS model in step 4 is expressed as:
combining the normalized feature matrix and daily vermicelli increment into a mixed matrix R m*n N is the number of features plus 1, m is the number of days of use of the user data; step by step
Step 4.1: the mixing matrix is approximately set to
Wherein X is m*k And
is R m*n The low-dimensional matrix obtained by decomposition, wherein the parameter k is the dimension of the low-dimensional matrix;
step 4.2: to make the two sides of the equation as equal as possible, the square error loss of the alternating least squares method is constructed
Loss function:
wherein R is ui 、X u 、Y i R is respectively m*n 、X m*k 、Y n*k L (X, Y) is a square error loss function;
step 4.3, at this point the matrix decomposition problem can be translated into a solution with minimum square error loss function:
the key to the ALS algorithm is to find the optimal X m*k And Y n*k Minimize L (X, Y) due to X m*k And Y n*k Is unknown, so the solution problem of the formula is not convex, the thought of the ALS algorithm is to fix one matrix and then solve the other matrix, and the specific algorithm is as follows:
step 4.3.1 firstly taking a random value to fix X u ；
Step 4.3.2 solving for Y for L (X, Y) i And let the bias guide be 0, can solve Y i
Y i ＝(x T x+λI) -1 x T R i (8)
Step 4.3.3 solving the above-mentioned Y i Fixed, X can be found by the same principle u
X u ＝(Y T Y+λI) -1 Y T R u (9)
Step 4.3.4 continuously repeats steps 4.3.2 and 4.3.3 until L (X, Y) reaches a target value or a maximum number of iterations is reached;
initially X u And Y i Is a random matrix, and the alternate least square method modifies X by continuous alternate iteration u And Y i Thereby obtaining the final X u And Y i Is a value of (2); solving for X u And Y i Reconstructing a mixed matrix through a formula 5, wherein the reconstructed mixed matrix contains the future expanding number of vermicelli to be solved;
step 5, optimizing training models of different data sources by using a genetic algorithm to obtain the dimension k of the optimal low-dimensional matrix;
and step 6, embedding the vermicelli expansion model obtained by training into a platform system, and applying the vermicelli expansion model practically.
2. The SFT-ALS-based time-series vermicelli expansion prediction method of claim 1, wherein: the genetic algorithm optimizing process in the step 5 is expressed as follows:
k＝GA(x′，ALS) (10)
where x' is the feature matrix of the dataset, ALS represents the ALS model trained by the dataset, and GA (·) is a genetic algorithm function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110188635.9A CN112862195B (en) | 2021-02-19 | 2021-02-19 | SFT-ALS-based time series vermicelli fluctuation prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110188635.9A CN112862195B (en) | 2021-02-19 | 2021-02-19 | SFT-ALS-based time series vermicelli fluctuation prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112862195A CN112862195A (en) | 2021-05-28 |
| CN112862195B true CN112862195B (en) | 2023-06-20 |
Family
ID=75988176
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110188635.9A Active CN112862195B (en) | 2021-02-19 | 2021-02-19 | SFT-ALS-based time series vermicelli fluctuation prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112862195B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114489156B (en) * | 2022-01-25 | 2024-05-07 | 南京晨光集团有限责任公司 | High-precision control strategy based on three-axis turntable |
| CN114756709A (en) * | 2022-03-25 | 2022-07-15 | 北京卓越乐享网络科技有限公司 | Comparison method, device, equipment, storage medium and program product for publisher |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573731A (en) * | 2015-02-06 | 2015-04-29 | 厦门大学 | Rapid target detection method based on convolutional neural network |
| CN105550275A (en) * | 2015-12-09 | 2016-05-04 | 中国科学院重庆绿色智能技术研究院 | Microblog forwarding quantity prediction method |
| CN106682118A (en) * | 2016-12-08 | 2017-05-17 | 华中科技大学 | Social network site false fan detection method achieved on basis of network crawler by means of machine learning |
| CN109558563A (en) * | 2017-09-26 | 2019-04-02 | 中国人民解放军信息工程大学 | Degree of rarefication adaptive sparse Fourier transform method, terminal and storage medium |
| CN109711865A (en) * | 2018-12-07 | 2019-05-03 | 恒安嘉新（北京）科技股份公司 | A method of prediction is refined based on the mobile radio communication flow that user behavior excavates |
| CN110276491A (en) * | 2019-06-24 | 2019-09-24 | 重庆锐云科技有限公司 | Bean vermicelli prediction management method, apparatus, computer equipment and storage medium |
| CN110325929A (en) * | 2016-12-07 | 2019-10-11 | 阿瑞路资讯安全科技股份有限公司 | System and method for detecting the waveform analysis of cable network variation |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11263543B2 (en) * | 2007-11-02 | 2022-03-01 | Ebay Inc. | Node bootstrapping in a social graph |
| CN102346745B (en) * | 2010-08-02 | 2014-04-02 | 阿里巴巴集团控股有限公司 | Method and device for predicting user behavior number for words |
| JP6831280B2 (en) * | 2017-03-24 | 2021-02-17 | 株式会社日立製作所 | Prediction system and prediction method |
-
2021
- 2021-02-19 CN CN202110188635.9A patent/CN112862195B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104573731A (en) * | 2015-02-06 | 2015-04-29 | 厦门大学 | Rapid target detection method based on convolutional neural network |
| CN105550275A (en) * | 2015-12-09 | 2016-05-04 | 中国科学院重庆绿色智能技术研究院 | Microblog forwarding quantity prediction method |
| CN110325929A (en) * | 2016-12-07 | 2019-10-11 | 阿瑞路资讯安全科技股份有限公司 | System and method for detecting the waveform analysis of cable network variation |
| CN106682118A (en) * | 2016-12-08 | 2017-05-17 | 华中科技大学 | Social network site false fan detection method achieved on basis of network crawler by means of machine learning |
| CN109558563A (en) * | 2017-09-26 | 2019-04-02 | 中国人民解放军信息工程大学 | Degree of rarefication adaptive sparse Fourier transform method, terminal and storage medium |
| CN109711865A (en) * | 2018-12-07 | 2019-05-03 | 恒安嘉新（北京）科技股份公司 | A method of prediction is refined based on the mobile radio communication flow that user behavior excavates |
| CN110276491A (en) * | 2019-06-24 | 2019-09-24 | 重庆锐云科技有限公司 | Bean vermicelli prediction management method, apparatus, computer equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 一种稀疏度自适应的稀疏傅里叶变换算法;刘仲 等;计算机工程;第第44卷卷(第第2期期);141-146 * |
| 基于改进DE-Tri-Training算法的汉语多词表达抽取;黄丹丹 等;数据采集与处理;第32卷(第1期);141-148 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112862195A (en) | 2021-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112862195B (en) | SFT-ALS-based time series vermicelli fluctuation prediction method | |
| Zeng et al. | Monotone-delay-interval-based Lyapunov functionals for stability analysis of systems with a periodically varying delay | |
| CN112926303B (en) | Malicious URL detection method based on BERT-BiGRU | |
| CN109992764A (en) | A kind of official documents and correspondence generation method and device | |
| Caspersen et al. | How stand productivity results from size-and competition-dependent growth and mortality | |
| CN109005451B (en) | Video strip splitting method based on deep learning | |
| Al-Saleem et al. | Mining educational data to predict students’ academic performance | |
| CN106780701B (en) | Non-uniform texture image synthesis control method, device, storage medium and equipment | |
| Broad et al. | Active Divergence with Generative Deep Learning--A Survey and Taxonomy | |
| Liu | Music trend prediction based on improved LSTM and random forest algorithm | |
| CN107027051A (en) | A kind of video key frame extracting method based on linear dynamic system | |
| Huang et al. | Diffstyler: Controllable dual diffusion for text-driven image stylization | |
| CN114780083A (en) | Visual construction method and device of knowledge map system | |
| Faraldo et al. | A multi-profile method for key estimation in EDM | |
| CN112182221A (en) | Knowledge retrieval optimization method based on improved random forest | |
| CN106600213A (en) | Intelligent resume management system and method | |
| Anderson et al. | EMPOWER-1.0: an efficient model of planktonic ecosystems written in R | |
| Kitagawa et al. | Split-based points from the Swabian Jura highlight Aurignacian regional signatures | |
| DE102022113243A1 (en) | Score-based generative modeling in latent space | |
| Hafner et al. | Perceptions of bioeconomy and the desire for governmental action: Regional actors’ connotations of wood-based bioeconomy in Germany | |
| Beyer et al. | Validation of a functional-structural tree model using terrestrial Lidar data | |
| CN113962417A (en) | Video processing method and device, electronic equipment and storage medium | |
| CN113962335B (en) | Flexibly configurable data whole-process processing method | |
| Piras et al. | Forest surface changes and cultural values: The forests of Tuscany (Italy) in the last century | |
| Kaliakatsos-Papakostas et al. | Controlling interactive evolution of 8-bit melodies with genetic programming |
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 |