CN109840551A - A method of the optimization random forest parameter for machine learning model training - Google Patents
A method of the optimization random forest parameter for machine learning model training Download PDFInfo
- Publication number
- CN109840551A CN109840551A CN201910030755.9A CN201910030755A CN109840551A CN 109840551 A CN109840551 A CN 109840551A CN 201910030755 A CN201910030755 A CN 201910030755A CN 109840551 A CN109840551 A CN 109840551A
- Authority
- CN
- China
- Prior art keywords
- ant
- population
- lion
- ant lion
- spark
- 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.)
- Granted
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of methods of optimization random forest parameter for machine learning model training, are first several sub- populations by entire ant population dividing；Then each sub- population is corresponded into a subregion in RDD, and specifies independent evolve in a subregion；Finally information is exchanged between each sub- population using transfer operator.Compared to traditional grid search, parallel ant lion algorithm based on Spark can find more excellent parameter combination efficiently to improve the nicety of grading of random forest, and under big data distribution Spark platform, optimizing speed is fast, acceleration effect is obvious, can be used as the next-generation parameter optimiser of cloud computing platform.
Description
Technical field
The invention belongs to large-scale machines learning model training fields, are related to a kind of for the excellent of machine learning model training
The method for changing random forest parameter is specifically related to a kind of parallel ant lion based on Spark for machine learning model training
The method of algorithm optimization random forest parameter.
Background technique
Recently as the sharp increase of information data, machine learning model has obtained extensive promotion and application.However
As a ring of model training most critical, the parameter optimization of model is always to compare stubborn problem, also tends to consume and compare
Long engineering time.Classic optimization method solution procedure will lead to bigger time complexity or space complexity, final to join
Number optimum level and the functional form of institute's Solve problems are closely related.Swarm intelligence algorithm is as a kind of mimic biology group behavior
Random search algorithm has been constantly subjected to widely pay close attention to and apply since proposition.Different population intelligent algorithm has different
Theoretical foundation and realization step, ant lion algorithm (The Ant Lion Optimizer, ALO) is Seyedali Mirjalili
In a kind of novel meta-heuristic swarm intelligence algorithm that 2015 propose.The inspiration of algorithm ant lion larva in the Nature
Foraging behavior, main be centered around around randomly selected ant lion by ant carry out random walk to realize global search.?
Foreign countries have been successfully applied to member structure optimization, the training of the distribution problem of idle work, multilayer neural network, have filtered to linear discrete
Wave device carries out the fields such as the flight course planning problem of High Efficiency Modeling and unmanned plane.
And random forest is not only realized simply as a kind of novel integrated classifier, but also can handle high dimensional data
And quickly obtain classification results.It needs the parameter of user setting relatively fewer in algorithm, has relative to other conventional sorting methods
There is good robustness, thus is received significant attention in classification problem.But in order to further increase nicety of grading and efficiency, lead to
It crosses and random forests algorithm key parameter is optimized, may be implemented in random forest sort run efficiency tolerance interval
More high-class precision.Therefore the parameter optimization of random forest classification becomes a feasible way for improving nicety of grading.
But since traditional intelligence algorithm is under single-processor environment, parallel search can only be realized in a serial fashion, it is this
The data set that mode increases under unsuitable big data era on a large scale.The rise of cloud computing technology is the parallel of ant lion algorithm
Change realization and provide a fine resolving ideas, merges the technologies such as parallel computation and distributed computing, and extensive
There is brilliant performance in terms of isomeric data storage and logic calculation.
Summary of the invention
In order to overcome the above-mentioned deficiencies of the prior art, the present invention is intended to provide a kind of base for machine learning model training
In the method for the parallel ant lion algorithm optimization random forest parameter of Spark.
Technical solution used by method of the invention is: a kind of optimization random forest for machine learning model training
The method of parameter, which comprises the following steps:
Step 1: by the storage of collected training set data into distributed file system, store path is variable
path；
Step 2: global optimization optimizing, the parameter group optimized are carried out using parameter of the ant lion algorithm to random forest
It closes；
Specific implementation includes following sub-step:
Step 2.1: on host node Spark Drive, initializing the information of entire population, including initial ant population number
Amount, initial ant lion population quantity and the maximum number of iterations of population；
Step 2.2: initialization SparkContext is variable sc, and reads data set using spark function textFile
Path path obtains distributed training dataset, is variables D ataRDD；
Step 2.3: being directed to initial ant lion population, calculate each ant lion individual fitness according to value function is adapted to, and select
Maximum adaptation value individual is elite ant lion primary；
Step 2.4: entire initial ant population is turned into distributed ant colony parallel using spark function parallelize,
For variable antRDD；
Step 2.5: by the sub- population of each ant to each of dependent variable antRDD subregion, and in each subregion
In specify independent evolve；
Step 2.6: being solved carrying out the sub- population optimizing of parallel ant from node Worker, complete entire ant population
Updating location information；
Step 3: the sub- population of the ant of each subregion is pooled to host node Spark Driver from node Worker from each
On merge, obtain updated next-generation ant population；
Step 4: the rule determined according to adaptive value compares ant lion population and ant on host node Spark Driver
The adaptive value of population, and then the update of entire ant lion population position is completed, the next-generation ant lion group after generating the survival of the fittest, together
When judge whether elite ant lion individual needs to update；
Step 5: the information of updated ant lion population and elite ant lion being re-broadcast each from node Worker；
Step 6: judging whether ant lion population iteration reaches maximum number of iterations；
If it is not, then returning to step 2.4；
If so, returning to last elite ant lion, the optimal solution that as current ant lion algorithm searches out.
The present invention using Spark cluster distributed arithmetic based on memory characteristic, ant population dividing at multiple ants
Sub- Species structure realizes concurrent operation in each section of cluster.Compared to traditional grid search, the parallel ant lion based on Spark
Algorithm can find more excellent parameter combination efficiently to improve the nicety of grading of random forest, and flat in big data distribution Spark
Under platform, optimizing speed is fast, and acceleration effect is obvious, can be used as the next-generation parameter optimiser of cloud computing platform.
Detailed description of the invention
Fig. 1 is the flow chart of the embodiment of the present invention；
Fig. 2 is the parallel search flow chart of the embodiment of the present invention；
Fig. 3 is the ant lion algorithm architecture diagram based on Spark of the embodiment of the present invention；
Fig. 4 is random forest schematic diagram in the ant lion algorithm based on Spark of the embodiment of the present invention.
Specific embodiment
Understand for the ease of those of ordinary skill in the art and implement the present invention, with reference to the accompanying drawings and embodiments to this hair
It is bright to be described in further detail, it should be understood that implementation example described herein is merely to illustrate and explain the present invention, not
For limiting the present invention.
See Fig. 1 and Fig. 2, a kind of optimization random forest parameter for machine learning model training provided by the invention
Method, which comprises the following steps:
Step 1: by the storage of collected training set data into HDFS distributed file system, store path is variable
Path, by taking one letter from home of the U.S. is with finance company LendingClub as an example, debt-credit data can be obtained from official website；
Step 2: global optimization optimizing, the parameter group optimized are carried out using parameter of the ant lion algorithm to random forest
It closes；
Specific implementation includes following sub-step:
Step 2.1: on host node Spark Drive, initializing the information of entire population, including initial ant population number
Amount, initial ant lion population quantity and the maximum number of iterations of population；
Step 2.2: initialization SparkContext is variable sc, and reads data set using spark function textFile
Path path obtains distributed training dataset, is variables D ataRDD；
Step 2.3: being directed to initial ant lion population, calculate each ant lion individual fitness according to value function is adapted to, and select
Maximum adaptation value individual is elite ant lion primary；
Specific implementation includes following sub-step:
Step 2.3.1: need to calculate the population pop of adaptive value to each from section using spark function broadcast broadcast
It is variable popbro on point Worker；
Step 2.3.2: on distributed data collection DataRDD, obtain current partition in all data characteristic value X and
Corresponding data label Y；
Step 2.3.3: whole from broadcast variable popbro acquisition in current partition on distributed data collection DataRDD
A population is variable pop；
Step 2.3.4: successively traversing population pop in each subregion, and the adaptation of each individual is calculated using random forest
Value；When calculating adaptive value, value function is adapted to using following:
Fitness=w1*OOBscore+w2*feature-1
Wherein: w1The weight of presentation class precision；OOBscoreThe OOB scoring divided for random forest character subset；w2Table
Show characteristic ratio coding weight reciprocal；Feature is characterized continuous coding；The meaning designed in this way is, parameter exist
It wishes to take into account the accuracy rate of model and the training speed of model when training pattern on large-scale dataset；
Step 2.3.6: it to each individual is directed to, is averaged, is made using adaptive value of the function reduce to all subregions
For the adaptive value of last individual；
Step 2.3.7: current ant lion fitness value is arranged according to descending, the ant lion information to rank the first is recorded primary
Elite ant lion；
Step 2.4: entire initial ant population is turned into distributed ant colony parallel using spark function parallelize,
For variable antRDD；
Specific implementation includes following sub-step:
Step 2.4.1: ant lion population primary and corresponding adaptive value are broadcasted using spark function broadcast；
Step 2.4.2: according to ant colony scale, creation creates each ant individual and numbers, and is [1,2,3,4 ... n]；
Ant colony: being turned to distributed ant population using spark function parallelize by step 2.4.3 parallel, is variable
antRDD。
Step 2.5: by the sub- population of each ant to a subregion in dependent variable antRDD, and in each subregion
Specified independent evolution；
Specific implementation includes following sub-step:
Step 2.5.1: spark function mapPartitionsWithIndex is utilized on distributed ant population antRDD
Carry out each Paralleled global optimizing；
Step 2.5.2: according to ant lion adaptive value on each subregion, predator ant lion RA is obtained by roulette；
Step 2.5.3: according to Archimedes spiral formula, simulate ant by ant lion along swirl shape convergence route gradually by
The scene of capture:
In formula:For the position of the ant lion jth dimension of t generation capture ant； For antWith its ant lion of captureAnd elite ant lionBetween weight distance and；w1And w2Point
Not Wei predator ant lion and elite ant lion distance influence factor；B is helical curve constant；T is when former generation the number of iterations；T is
Maximum number of iterations；Random number of the rand between [0,1]；T ' ∈ [- 2,1], and change with the number of iterations t；
Step 2.5.4: it updated result will be returned from each subregion of node Worker using spark function collectc
Host node Spark Driver is returned to, each division result includes key-value pair (partitionIndex, Antpop), wherein
PartitionIndex is partition number, and Antpop is updated sub- population；
Step 2.5.5: merge the sub- population of ant, compare ant lion population and ant population on host node Spark Driver
Adaptive value, complete ant lion population position update.
Step 2.6: being solved carrying out the sub- population optimizing of parallel ant from node Worker, complete entire ant population
Updating location information；
Step 3: the sub- population of the ant of each subregion is pooled to host node Spark Driver from node Worker from each
On merge, obtain updated next-generation ant population；
Step 4: the rule determined according to adaptive value compares ant lion population and ant on host node Spark Driver
The adaptive value of population, and then the update of entire ant lion population position is completed, the next-generation ant lion group after generating the survival of the fittest, together
When judge whether elite ant lion individual needs to update；
Step 5: the information of updated ant lion population and elite ant lion being re-broadcast each from node Worker；
Step 6: judging whether ant lion population iteration reaches maximum number of iterations；
If it is not, then returning to step 2.4；
If so, returning to last elite ant lion, the optimal solution that as current ant lion algorithm searches out.
It is the ant lion algorithm architecture diagram based on Spark of the present embodiment see Fig. 3.First layer is data needed for calculating
Acquisition；The second layer is the large-scale distributed file storage service that Hadoop HDFS is provided；Third layer contains Spark most core
The function of the heart and distributed operator, wherein covering query engine Spark SQL for supporting structuring data query and analysis, distribution
Formula machine learning library Spark MLlib, parallel figure Computational frame Graph X, stream calculation frame Spark tetra- moulds of Streaming
Block；4th layer is then ant lion algorithm for random forest progress hyper parameter optimizing, and each sub- population parallel search exchanges information, finally
Complete the process that optimal ant lion finds.
It is random forest schematic diagram in the ant lion algorithm based on Spark of the present embodiment, first from input see Fig. 4
The sampling put back to is taken in training set, constructs Sub Data Set.Then single sub-tree is constructed using Sub Data Set.Finally
It needs to obtain classification results by random forest for new test data, according to the judging result of sub-tree, selection occurs
The most classification of number is the output result of final random forest.
It should be understood that the part that this specification does not elaborate belongs to the prior art.
It should be understood that the above-mentioned description for preferred embodiment is more detailed, can not therefore be considered to this
The limitation of invention patent protection range, those skilled in the art under the inspiration of the present invention, are not departing from power of the present invention
Benefit requires to make replacement or deformation under protected ambit, fall within the scope of protection of the present invention, this hair
It is bright range is claimed to be determined by the appended claims.
Claims (4)
1. a kind of method of the optimization random forest parameter for machine learning model training, which is characterized in that including following step
It is rapid:
Step 1: by the storage of collected training set data into distributed file system, store path is variable path；
Step 2: global optimization optimizing, the parameter combination optimized are carried out using parameter of the ant lion algorithm to random forest；
Specific implementation includes following sub-step:
Step 2.1: on host node Spark Drive, initialize the information of entire population, including initial ant population quantity,
Initial ant lion population quantity and the maximum number of iterations of population；
Step 2.2: initialization SparkContext is variable sc, and reads data set path using spark function textFile
Path obtains distributed training dataset, is variables D ataRDD；
Step 2.3: being directed to initial ant lion population, calculate each ant lion individual fitness according to value function is adapted to, and select maximum
Adaptive value individual is elite ant lion primary；
Step 2.4: entire initial ant population is turned into distributed ant population parallel using spark function parallelize,
For variable antRDD；
Step 2.5: by the sub- population of each ant to each of dependent variable antRDD subregion, and referring in each subregion
Fixed independent evolution；
Step 2.6: being solved carrying out the sub- population optimizing of parallel ant from node Worker, complete the position of entire ant population
Information update；
Step 3: by the sub- population of the ant of each subregion, from node Worker to be pooled to host node Spark Driver enterprising from each
Row merges, and obtains updated next-generation ant population；
Step 4: the rule determined according to adaptive value compares ant lion population and ant population on host node Spark Driver
Adaptive value, and then complete entire ant lion population position update, generate the survival of the fittest after next-generation ant lion group, sentence simultaneously
Whether disconnected elite ant lion individual needs to update；
Step 5: the information of updated ant lion population and elite ant lion being re-broadcast each from node Worker；
Step 6: judging whether ant lion population iteration reaches maximum number of iterations；
If it is not, then returning to step 2.4；
If so, returning to last elite ant lion, the optimal solution that as current ant lion algorithm searches out.
2. the method for the optimization random forest parameter according to claim 1 for machine learning model training, feature
It is, the specific implementation of step 2.3 includes following sub-step:
Step 2.3.1: need to calculate the population pop of adaptive value to each from node using spark function broadcast broadcast
It is variable popbro on Worker；
Step 2.3.2: on distributed data collection DataRDD, the characteristic value X and correspondence of all data in current partition are obtained
Data label Y；
Step 2.3.3: on distributed data collection DataRDD, entire kind is obtained from broadcast variable popbro in current partition
Group is variable pop；
Step 2.3.4: successively traversing population variable pop in each subregion, and the adaptation of each individual is calculated using random forest
Value；When calculating adaptive value, value function is adapted to using following:
Fitness=w1*OOBscore+w2*feature-1
Wherein: w1The weight of presentation class precision；OOBscoreThe OOB scoring divided for random forest character subset；w2Indicate feature
The weight of continuous coding inverse；Feature is characterized continuous coding；
Step 2.3.6: to each individual is directed to, function red is utilizeduceThe adaptive value of all subregions is averaged, as last
The adaptive value of individual；
Step 2.3.7: current ant lion fitness value is arranged according to descending, the ant lion information to rank the first is recorded into elite primary
Ant lion.
3. the method for the optimization random forest parameter according to claim 1 for machine learning model training, feature
It is, the specific implementation of step 2.4 includes following sub-step:
Step 2.4.1: ant lion population primary and corresponding adaptive value are broadcasted using spark function broadcast；
Step 2.4.2: according to ant colony scale, creation creates each ant individual and numbers, and is [1,2,3,4...n]；
Ant colony: being turned to distributed ant population using spark function parallelize by step 2.4.3 parallel, is variable
antRDD。
4. the method for the optimization random forest parameter according to claim 1 for machine learning model training, feature
It is, the specific implementation of step 2.5 includes following sub-step:
Step 2.5.1: it is carried out on distributed ant population antRDD using spark function mapPartitionsWithIndex
Each Paralleled global optimizing；
Step 2.5.2: according to ant lion adaptive value on each subregion, predator ant lion RA is obtained by roulette；
Step 2.5.3: it according to Archimedes spiral formula, simulates ant and is gradually captured by ant lion along swirl shape convergence route
Scene:
In formula:For the position of the ant lion jth dimension of t generation capture ant； For antWith its ant lion of captureAnd elite ant lionBetween weight distance and；w1And w2
The respectively distance influence factor of predator ant lion and elite ant lion；B is helical curve constant；T is when former generation the number of iterations；T
For maximum number of iterations；Random number of the rand between [0,1]；T ' ∈ [- 2,1], and change with the number of iterations t；
Step 2.5.4: it updated result will be returned to from each subregion of node Worker using spark function collectc
Host node Spark Driver, each division result include key-value pair (partitionIndex, Antpop), wherein
PartitionIndex is partition number, and Antpop is updated sub- population；
Step 2.5.5: merging the sub- population of ant, and the suitable of ant lion population and ant population is compared on host node Spark Driver
It should be worth, complete the update of ant lion population position.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910030755.9A CN109840551B (en) | 2019-01-14 | 2019-01-14 | Method for optimizing random forest parameters for machine learning model training |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910030755.9A CN109840551B (en) | 2019-01-14 | 2019-01-14 | Method for optimizing random forest parameters for machine learning model training |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109840551A true CN109840551A (en) | 2019-06-04 |
| CN109840551B CN109840551B (en) | 2022-03-15 |
Family
ID=66883810
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910030755.9A Active CN109840551B (en) | 2019-01-14 | 2019-01-14 | Method for optimizing random forest parameters for machine learning model training |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109840551B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110867224A (en) * | 2019-10-28 | 2020-03-06 | 南通大学 | Multi-granularity Spark super-trust fuzzy method for large-scale brain pathology segmentation |
| CN111079935A (en) * | 2019-10-16 | 2020-04-28 | 重庆邮电大学 | Machine learning rapid large-scale sample signature method under spark |
| CN112216341A (en) * | 2020-09-16 | 2021-01-12 | 中国人民解放军国防科技大学 | Group behavior logic optimization method and computer readable storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104239501A (en) * | 2014-09-10 | 2014-12-24 | 中国电子科技集团公司第二十八研究所 | Mass video semantic annotation method based on Spark |
| CN105550374A (en) * | 2016-01-29 | 2016-05-04 | 湖南大学 | Random forest parallelization machine studying method for big data in Spark cloud service environment |
| CN106056134A (en) * | 2016-05-20 | 2016-10-26 | 重庆大学 | Semi-supervised random forests classification method based on Spark |
| CN108154174A (en) * | 2017-12-22 | 2018-06-12 | 北京工业大学 | A kind of random forest categorizing system based on core extreme learning machine and parallelization |
| CN109142946A (en) * | 2018-06-29 | 2019-01-04 | 东华大学 | Transformer fault detection method based on ant group algorithm optimization random forest |
-
2019
- 2019-01-14 CN CN201910030755.9A patent/CN109840551B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104239501A (en) * | 2014-09-10 | 2014-12-24 | 中国电子科技集团公司第二十八研究所 | Mass video semantic annotation method based on Spark |
| CN105550374A (en) * | 2016-01-29 | 2016-05-04 | 湖南大学 | Random forest parallelization machine studying method for big data in Spark cloud service environment |
| CN106056134A (en) * | 2016-05-20 | 2016-10-26 | 重庆大学 | Semi-supervised random forests classification method based on Spark |
| CN108154174A (en) * | 2017-12-22 | 2018-06-12 | 北京工业大学 | A kind of random forest categorizing system based on core extreme learning machine and parallelization |
| CN109142946A (en) * | 2018-06-29 | 2019-01-04 | 东华大学 | Transformer fault detection method based on ant group algorithm optimization random forest |
Non-Patent Citations (2)
| Title |
|---|
| DONG GAIFANG 等: "Cooperative ant colony-genetic algorithm based on spark", 《ELSEVIER》 * |
| 闫 春 等: "基于蚁群算法优化随机森林模型的汽车保险欺诈识别研究", 《保险研究 》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079935A (en) * | 2019-10-16 | 2020-04-28 | 重庆邮电大学 | Machine learning rapid large-scale sample signature method under spark |
| CN111079935B (en) * | 2019-10-16 | 2022-10-18 | 重庆邮电大学 | Machine learning rapid large-scale sample signature method under spark |
| CN110867224A (en) * | 2019-10-28 | 2020-03-06 | 南通大学 | Multi-granularity Spark super-trust fuzzy method for large-scale brain pathology segmentation |
| CN110867224B (en) * | 2019-10-28 | 2022-02-08 | 南通大学 | Multi-granularity Spark super-trust fuzzy method for large-scale brain pathology segmentation |
| CN112216341A (en) * | 2020-09-16 | 2021-01-12 | 中国人民解放军国防科技大学 | Group behavior logic optimization method and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109840551B (en) | 2022-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tsai et al. | Parallel cat swarm optimization | |
| WO2021109464A1 (en) | Personalized teaching resource recommendation method for large-scale users | |
| Nadimi-Shahraki et al. | A systematic review of the whale optimization algorithm: theoretical foundation, improvements, and hybridizations | |
| Zabihi et al. | A novel history-driven artificial bee colony algorithm for data clustering | |
| Moayedikia et al. | Weighted bee colony algorithm for discrete optimization problems with application to feature selection | |
| CN109840551A (en) | A method of the optimization random forest parameter for machine learning model training | |
| Mu et al. | Multi-objective ant colony optimization algorithm based on decomposition for community detection in complex networks | |
| CN109784405A (en) | Cross-module state search method and system based on pseudo label study and semantic consistency | |
| CN111461286B (en) | Spark parameter automatic optimization system and method based on evolutionary neural network | |
| Zou et al. | Reinforcement learning to diversify top-n recommendation | |
| CN110020435B (en) | Method for optimizing text feature selection by adopting parallel binary bat algorithm | |
| Yang et al. | Adaptive chaotic spherical evolution algorithm | |
| CN110046713A (en) | Robustness sequence learning method and its application based on multi-objective particle swarm optimization | |
| CN113255873A (en) | Clustering longicorn herd optimization method, system, computer equipment and storage medium | |
| CN106407379A (en) | Hadoop platform based movie recommendation method | |
| Hashemi et al. | A new particle swarm optimization algorithm for optimizing big data clustering | |
| Wang et al. | A carnivorous plant algorithm with heuristic decoding method for traveling salesman problem | |
| Chen et al. | A Spark-based Ant Lion algorithm for parameters optimization of random forest in credit classification | |
| Feng | Self-generation fuzzy modeling systems through hierarchical recursive-based particle swarm optimization | |
| CN114742593A (en) | Logistics storage center optimal site selection method and system | |
| CN107423438A (en) | The problem of one kind is based on PGM sorting technique | |
| Rizman Žalik et al. | Node attraction-facilitated evolution algorithm for community detection in networks | |
| Wang et al. | Robot Path Planning with Low Learning Cost Using a Novel K-means-based Pointer Networks | |
| CN117556147B (en) | Electronic commerce data classification recommendation system and method | |
| Pei | Hybrid immune clonal particle swarm optimization multi-objective algorithm for constrained optimization problems |
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 | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220524 Address after: 430000 Long'an, No. 38, MINYUAN Road, East Lake New Technology Development Zone, Wuhan City, Hubei Province Room 02, floor 11, unit a, Ganghui City Patentee after: Wuhan AGCO Software Technology Co.,Ltd. Address before: 430068 1, Lijia 1 village, Nanhu, Wuchang District, Wuhan, Hubei Patentee before: HUBEI University OF TECHNOLOGY |