CN112733935A - Fishing boat fishing mode prediction method based on Stacking algorithm - Google Patents
Fishing boat fishing mode prediction method based on Stacking algorithm Download PDFInfo
- Publication number
- CN112733935A CN112733935A CN202110026309.8A CN202110026309A CN112733935A CN 112733935 A CN112733935 A CN 112733935A CN 202110026309 A CN202110026309 A CN 202110026309A CN 112733935 A CN112733935 A CN 112733935A
- Authority
- CN
- China
- Prior art keywords
- fishing
- data
- fishing boat
- model
- algorithm
- 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.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (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)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Physics (AREA)
- Business, Economics & Management (AREA)
- Software Systems (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computing Systems (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Algebra (AREA)
- Game Theory and Decision Science (AREA)
- Databases & Information Systems (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
Abstract
The invention discloses a fishing boat fishing mode prediction method based on a Stacking algorithm, which comprises the following steps: preprocessing fishing boat ID, longitude and latitude coordinates, speed, course, reporting time and other information provided by a Beidou satellite in real time, and then performing characteristic engineering processing on the preprocessed data to obtain a plurality of characteristic vectors; training by using the plurality of feature vectors as the input of a machine learning model, and then predicting the fishing mode of the fishing boat; and finally, the model is accessed into a prediction system, so that the off-line and real-time prediction analysis of the fishing boat fishing mode can be realized. Compared with the traditional characteristic engineering, the invention provides a vector coding scheme based on the track sequence, which is helpful for improving the prediction precision; compared with a single model, the fishing boat fishing mode prediction model based on the Stacking algorithm can further improve the classification accuracy, is higher in generalization and better in stability, and can be applied to accurate judgment of large-batch marine fishing boat fishing modes.
Description
Technical Field
The invention relates to the field of ocean data mining technology and computer science, in particular to a fishing boat fishing mode prediction method based on a Stacking algorithm.
Background
The marine fishery production is a core plate of marine economy, and a fishing boat is used as an important fishing tool for the marine fishery production, and the scale of the fishing boat is continuously increased. Trawling, seining and trawling are the three most common fishing modes of marine fishing boats, and encompass fishery products on all layers of the sea. With the rapid growth of the number of fishing boats and the rapid development of science and technology, the development of marine fishery in China also faces a series of problems: (1) fishery resources in the near sea area are continuously declining; (2) the country develops the 'fallow period' and 'forbidden period', which causes the income of fishermen to be reduced; (3) deep sea trawl fishing causes great damage to seabed ecological environment and system; (4) fishing boat accidents still occur; (5) the international marine fishery resource development competition is intensified; (6) the sea rights dispute is exacerbated. Therefore, a method for effectively identifying the fishing mode of the fishing boat is urgently needed by a fishing boat supervision department, so that the marine fishing boat can be effectively supervised, and the method has very important practical significance for solving the problems of sustainable development of marine fishery, fisherman life safety, marine international dispute and the like in China.
The existing method is mainly based on the traditional manual supervision method or the navigation log extraction method to judge the fishing mode of the fishing boat, and has the following problems: the Beidou satellite equipment track data does not contain state data of fishing boat catching modes, and the data of the fishing boat catching modes are only extracted from navigation log data. There are two disadvantages to log data: on one hand, the log is manually input in the fishing process, and some errors exist; on the other hand, log data is usually generated only after the end of the fishing process, with large time intervals; and secondly, fishing boat data provided by the Beidou satellite are not fully mined and utilized at present.
Therefore, the fishing boat Beidou satellite equipment data is used, training is carried out through a Stacking model, and a marine fishing boat fishing mode prediction system is developed, so that the method has important practical significance for improving fishery management in China.
Disclosure of Invention
Aiming at the above description, the existing manual fishing boat fishing mode judgment has errors, and the existing single model method has obvious limitation in fishing boat fishing mode prediction.
The invention can be solved by the following technical scheme:
a fishing boat fishing mode prediction method based on a Stacking algorithm comprises the following steps:
a) carrying out data preprocessing on the provided Beidou fishing boat position data, carrying out missing value inspection and processing on sample data, then carrying out abnormal value processing, and finally carrying out data smoothing processing to obtain a more standard data set for machine learning;
b) performing feature engineering operation, including feature construction and feature selection, so as to construct a reasonable feature vector;
c) leading the selected feature vectors into a Stacking model for training to obtain a classification model capable of identifying the fishing mode of each fishing boat;
d) a visual prediction system is set up, and the model is accessed into the system, so that the off-line and real-time prediction analysis of the fishing mode of the fishing boat can be realized.
Further, the data preprocessing comprises: the missing value of the sample data is checked and processed, meanwhile, the data visualization analysis method is used for better visualization operation of the sample data, and redundant and wrong data are removed; then, detecting and deleting abnormal points of the fishing boat track by using a DBSCAN algorithm; and finally, carrying out data smoothing processing on the fishing boat track by using a median filtering algorithm.
Further, the DBSCAN algorithm is a density-based spatial clustering algorithm, that is, the number of objects contained in a certain area in a space is required to be not less than a certain specified threshold;
the flow of the DBSCAN algorithm is as follows: scanning the whole data set, checking an Eps neighborhood of each point to search a cluster, and if the Eps neighborhood of the point p contains more points than a threshold Minpts, creating a cluster taking p as a core object; DBSCAN then iteratively aggregates objects that are directly density reachable from these core objects, a process that may involve the merging of some density reachable clusters; finally, when no new points are added to any cluster, the process ends, wherein no data points contained in any cluster constitute outliers;
the Eps neighborhood: in the data set D, for any data point PiE D, taking the e D as the center of the circle, and the distance is less than or equal to the set of all points of Eps, wherein the formula is as follows:
Eps(p)＝{q∈D|dist(p,q)≤eps} (1)
the threshold value Minpts: in the data set D, for any data point PiAnd e, determining the quantity of data points in the Eps neighborhood as k according to the element belonging to the element.
Further, the median filtering algorithm is a denoising algorithm, and the values of one point in the trajectory are replaced by the median of the point values around the point, so that the values are close to eliminate the noise point in the original trajectory data.
Further, the feature vector comprises one or more of the following feature vectors:
a) location-based feature engineering;
b) speed-based feature engineering;
c) direction-based feature engineering;
d) and (4) carrying out feature engineering based on the word2vec model.
Further, the Stacking model has a two-layer architecture: the first layer combines different primary learners, including random forests, XGboost, LightGBM and GBDT; the second layer uses SVM as a secondary learner; and the second-layer learner uses the predicted result of the first layer as a characteristic and predicts the final result, and five-fold cross validation is used for reducing overfitting in the model construction process.
Further, the cross-validation is a practical method to statistically cut data samples into smaller subsets to prevent overfitting due to models being too complex.
Further, the fishing boat fishing mode comprises the following steps: purse seine fishing, gill net fishing and trawl fishing.
Further, the prediction system comprises: the fishing mode of the fishing boat is predicted through the algorithm model, and the fishing mode and the position track information of the fishing boat are displayed on a webpage in real time, so that the fishing boat fishing mode can be effectively monitored.
Advantageous effects
The invention provides a vector coding scheme based on a track sequence, which is helpful for improving the prediction precision; compared with a single model, the fishing boat fishing mode prediction model based on the Stacking algorithm can further improve the classification accuracy, is higher in generalization and better in stability, and can be applied to accurate judgment of large-batch marine fishing boat fishing modes.
Drawings
FIG. 1 is an overall flow framework of a fishing vessel fishing mode prediction model according to the present invention;
FIG. 2 is a track sequence based vector encoding scheme of the present invention;
FIG. 3 is the overall architecture of the Stacking model in the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification.
The invention discloses a fishing boat fishing mode prediction method based on a Stacking algorithm, and an overall flow frame is shown in figure 1. Based on fishing boat Beidou satellite equipment data, various machine learning technologies are combined by using a Stacking algorithm to construct a new prediction model, so that the fishing boat fishing mode can be predicted on line, and the accuracy of judging the fishing boat fishing mode is effectively improved. Through the visual prediction system, the convenience of fishing boat supervision personnel on monitoring the fishing boat is further improved.
The specific steps of the invention when implemented are as follows:
the method comprises the following steps: and (4) preprocessing data. The original fishing boat Beidou satellite equipment has complex data, and not only can the phenomenon of data loss exist, but also the situation of data abnormity can occur. Therefore, the original fishing boat data needs to be effectively preprocessed, so that the processed data meets the requirement of model learning, and the training effect of the model is improved. The pretreatment process comprises the following steps:
first, the raw fishing boat data is checked for missing values.
And secondly, detecting and deleting obvious abnormal outlier track points in the fishing boat track based on the DBSCAN algorithm. The flow of the DBSCAN algorithm is as follows: scanning the whole data set, checking an Eps neighborhood of each point to search a cluster, and if the Eps neighborhood of the point p contains more points than a threshold Minpts, creating a cluster taking p as a core object; DBSCAN then iteratively aggregates objects that are directly density reachable from these core objects, a process that may involve the merging of some density reachable clusters; finally, when no new point is added to any cluster, the process ends. Where data points not contained in any cluster constitute outliers.
Eps neighborhood: in the data set D, for any data point PiE D, and a set of all points with the distance less than or equal to Eps by taking the e D as a center of a circle. The formula is as follows:
Eps(p)＝{q∈D|dist(p,q)≤eps} (1)
threshold value Minpts: in the data set D, for any data point PiE D, the number of data points set in the neighborhood of Eps is k. When k is larger than or equal to Minpts, the Eps neighborhood is classified into a class, wherein Minpts is the threshold of the neighborhood.
And finally, smoothing the fishing boat track data based on median filtering. The median filtering technology can effectively inhibit noise, and the values of one point in the track are replaced by the median of the point values around the point, so that the values are close to each other to eliminate the noise point in the original track data.
Step two: and the characteristic engineering is used for operating the preprocessed data based on the characteristic construction and the characteristic selection.
The original Beidou satellite equipment data set only has 5 original fields such as longitude and latitude, speed, course, reporting time and the like, can be directly used with few characteristics, and has an unsatisfactory effect of predicting fishing modes of fishing boats. Therefore, on the basis of the original features, the feature derivation is carried out on the original features by applying a statistical analysis method. For example, feature derivation based on position, statistics of mean, variance, median, quantile, skewness, kurtosis and the like of longitude and latitude are performed; carrying out characteristic derivation based on the speed, and counting the mean value, variance, skewness, kurtosis and the like of the speed; and (4) carrying out direction-based feature derivation, and counting the first-order difference, the variance and the like of the direction. Since the track characteristics of the three fishing boat fishing modes are different, in addition to the conventional statistical characteristics, the invention also uses a vector coding scheme (figure 2) based on track sequences aiming at the track characteristics of different fishing modes (purse net fishing, gill net fishing or trawl fishing) of the fishing boats. The text vector training word2vec model is used for calculating embedding characteristics of each position so as to be used for training a later stage Stackig model, and model prediction precision is improved.
The finally constructed features may have high correlation, which is not beneficial to the training of the model. Therefore, correlation analysis needs to be performed on all the constructed features, and the variables with high correlation are eliminated, so that the speed and the accuracy of model training are improved. When the feature selection is carried out, the selected feature selection method is a Pearson correlation coefficient, which is a simple and good feature selection method, and the feature selection effect is also good.
The Pearson correlation coefficient has a value between-1 and 1, can reflect the linear correlation between the characteristic and the predicted value, and describes the trend of simultaneous change of two groups of linear data. Wherein if the feature and the predicted value are absolute negative correlations, the Pearson value is-1; if the characteristic and the predicted value do not have a linear correlation relationship, the Pearson value is 0; if the feature and the predicted value are in absolute positive correlation, the Pearson value is 1. The calculation formula is as follows:
cov (X, Y) in equation (2) represents the covariance of two variables, which is calculated as follows:
in equation (2) (. sigma)XThe standard deviation of the variable X is expressed by the following formula:
in equation (2) (. sigma)YThe standard deviation of the variable Y is expressed by the following formula:
step three: and constructing a prediction model based on a Stacking algorithm. Using a Stacking mode, establishing a two-layer framework: the first layer combines different primary learners, including random forests, XGboost, LightGBM and GBDT; the second layer uses SVMs as secondary learners. The second-tier learner uses the results of the first-tier prediction as features and predicts the final results. In the model construction process, five-fold cross validation is used in order to reduce overfitting. The general architecture of the Stacking model is shown in fig. 3.
Step four: and (5) training and verifying the model. And randomly cutting 70% of data from the fishing boat sample data set to be used as a training set for model construction, and taking the rest 30% of data as a test set for model inspection. The model performance was measured using Accuracy (Accuracy), Precision (Precision), Recall (Recall), and 3 categories of respective F1 value averages (Score).
Step five: provided is a marine fishing boat fishing mode prediction system. A marine fishing boat fishing mode prediction system based on a Web front end is built, and the fishing mode of a fishing boat can be visually analyzed by accessing a Stacking model into the system.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (9)
1. A fishing boat fishing mode prediction method based on a Stacking algorithm is characterized by comprising the following steps: the method comprises the following steps:
a) carrying out data preprocessing on the provided Beidou fishing boat position data, carrying out missing value inspection and processing on sample data, then carrying out abnormal value processing, and finally carrying out data smoothing processing to obtain a more standard data set for machine learning;
b) performing feature engineering operation, including feature construction and feature selection, so as to construct a reasonable feature vector;
c) leading the selected feature vectors into a Stacking model for training to obtain a classification model capable of identifying the fishing mode of each fishing boat;
d) a visual prediction system is set up, and the model is accessed into the system, so that the off-line and real-time prediction analysis of the fishing mode of the fishing boat can be realized.
2. The method for predicting fishing vessel catching mode based on Stacking algorithm according to claim 1, wherein the data preprocessing comprises: the missing value of the sample data is checked and processed, meanwhile, the data visualization analysis method is used for better visualization operation of the sample data, and redundant and wrong data are removed; then, detecting and deleting abnormal points of the fishing boat track by using a DBSCAN algorithm; and finally, carrying out data smoothing processing on the fishing boat track by using a median filtering algorithm.
3. The fishing vessel fishing mode prediction method based on the Stacking algorithm according to claim 2, wherein the DBSCAN algorithm is a density-based spatial clustering algorithm, that is, the number of objects contained in a certain area in a space is required to be not less than a certain specified threshold;
the flow of the DBSCAN algorithm is as follows: scanning the whole data set, checking an Eps neighborhood of each point to search a cluster, and if the Eps neighborhood of the point p contains more points than a threshold Minpts, creating a cluster taking p as a core object; DBSCAN then iteratively aggregates objects that are directly density reachable from these core objects, a process that may involve the merging of some density reachable clusters; finally, when no new points are added to any cluster, the process ends, wherein no data points contained in any cluster constitute outliers;
the Eps neighborhood: in the data set D, for any data point PiE D, taking the e D as the center of the circle, and the distance is less than or equal to the set of all points of Eps, wherein the formula is as follows:
Eps(p)＝{q∈D|dist(p,q)≤eps}
the threshold value Minpts: in the data set D, for any data point PiAnd e, determining the quantity of data points in the Eps neighborhood as k according to the element belonging to the element.
4. The method for predicting fishing vessel fishing modes based on Stacking algorithm according to claim 2, wherein the median filtering algorithm is a denoising algorithm, and the values of one point in the trajectory are replaced by median of the point values around the point, so that the values are approximated to eliminate the noise point in the original trajectory data.
5. The method for predicting fishing vessel catching mode based on Stacking algorithm according to claim 1, wherein the feature vector comprises one or more of the following feature vectors:
a) location-based feature engineering;
b) speed-based feature engineering;
c) direction-based feature engineering;
d) and (4) carrying out feature engineering based on the word2vec model.
6. The method for predicting fishing vessel catching mode based on Stacking algorithm according to claim 1, wherein the Stacking model is of a two-layer structure: the first layer combines different primary learners, including random forests, XGboost, LightGBM and GBDT; the second layer uses SVM as a secondary learner; and the second-layer learner uses the predicted result of the first layer as a characteristic and predicts the final result, and five-fold cross validation is used for reducing overfitting in the model construction process.
7. The method for predicting fishing vessel fishing patterns based on the Stacking algorithm as claimed in claim 6, wherein the cross validation is a practical method for statistically cutting data samples into smaller subsets for preventing overfitting caused by too complicated model.
8. The method for predicting fishing vessel catching mode based on Stacking algorithm according to claim 1, wherein the fishing vessel catching mode comprises: purse seine fishing, gill net fishing and trawl fishing.
9. The method for predicting fishing vessel catching mode based on Stacking algorithm according to claim 1, wherein the prediction system comprises: the fishing mode of the fishing boat is predicted through the algorithm model, and the fishing mode and the position track information of the fishing boat are displayed on a webpage in real time, so that the fishing boat fishing mode can be effectively monitored.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110026309.8A CN112733935A (en) | 2021-01-08 | 2021-01-08 | Fishing boat fishing mode prediction method based on Stacking algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110026309.8A CN112733935A (en) | 2021-01-08 | 2021-01-08 | Fishing boat fishing mode prediction method based on Stacking algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112733935A true CN112733935A (en) | 2021-04-30 |
Family
ID=75591364
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110026309.8A Withdrawn CN112733935A (en) | 2021-01-08 | 2021-01-08 | Fishing boat fishing mode prediction method based on Stacking algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112733935A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113298195A (en) * | 2021-07-27 | 2021-08-24 | 中国电子科技集团公司第十五研究所 | Method and device for generating classical trajectory of offshore target and storage medium |
| CN117592818A (en) * | 2024-01-17 | 2024-02-23 | 武汉理工大学 | Auxiliary decision-making method and device for illegal fishing event |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110175633A (en) * | 2019-04-29 | 2019-08-27 | 杭州电子科技大学 | A kind of fishing boat track method for detecting abnormality based on Beidou satellite transmission data |
| CN110309383A (en) * | 2019-06-17 | 2019-10-08 | 武汉科技大学 | Ship trajectory clustering analysis method based on improved DBSCAN algorithm |
-
2021
- 2021-01-08 CN CN202110026309.8A patent/CN112733935A/en not_active Withdrawn
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110175633A (en) * | 2019-04-29 | 2019-08-27 | 杭州电子科技大学 | A kind of fishing boat track method for detecting abnormality based on Beidou satellite transmission data |
| CN110309383A (en) * | 2019-06-17 | 2019-10-08 | 武汉科技大学 | Ship trajectory clustering analysis method based on improved DBSCAN algorithm |
Non-Patent Citations (4)
| Title |
|---|
| GIULIANA PALLOTTA ET AL.: "Vessel Pattern Knowledge Discovery from AIS Data: A Framework for Anomaly Detection and Route Prediction", 《ENTROPY》 * |
| 李天举等: "基于集成学习的烟草异常数据挖掘研究与应用", 《计算机技术与发展》 * |
| 郑巧玲等: "基于神经网络和VMS的渔船捕捞类型辨别", 《南方水产科学》 * |
| 陈宏昆等: "基于船舶自动识别系统信息和Hough变换的海上船舶航道提取", 《计算机应用》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113298195A (en) * | 2021-07-27 | 2021-08-24 | 中国电子科技集团公司第十五研究所 | Method and device for generating classical trajectory of offshore target and storage medium |
| CN117592818A (en) * | 2024-01-17 | 2024-02-23 | 武汉理工大学 | Auxiliary decision-making method and device for illegal fishing event |
| CN117592818B (en) * | 2024-01-17 | 2024-04-09 | 武汉理工大学 | Auxiliary decision-making method and device for illegal fishing event |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Gao et al. | Ship-handling behavior pattern recognition using AIS sub-trajectory clustering analysis based on the T-SNE and spectral clustering algorithms | |
| Zhang et al. | A predictive analytics method for maritime traffic flow complexity estimation in inland waterways | |
| CN108023876B (en) | Intrusion detection method and intrusion detection system based on sustainability ensemble learning | |
| Winker et al. | Proof of concept for a novel procedure to standardize multispecies catch and effort data | |
| Li et al. | AIS data-driven ship trajectory prediction modelling and analysis based on machine learning and deep learning methods | |
| CN110796270A (en) | Machine learning model selection method | |
| CN109753591A (en) | Operation flow predictability monitoring method | |
| Duan et al. | A semi-supervised deep learning approach for vessel trajectory classification based on AIS data | |
| CN112733935A (en) | Fishing boat fishing mode prediction method based on Stacking algorithm | |
| CN110610165A (en) | Ship behavior analysis method based on YOLO model | |
| WO2019200739A1 (en) | Data fraud identification method, apparatus, computer device, and storage medium | |
| CN112101574B (en) | Machine learning supervised model interpretation method, system and equipment | |
| Zhang et al. | An automatically learning and discovering human fishing behaviors scheme for CPSCN | |
| CN116132104A (en) | Intrusion detection method, system, equipment and medium based on improved CNN-LSTM | |
| CN115130600A (en) | High-dimensional dynamic data stream anomaly detection method based on stacking habituation self-encoder | |
| CN113626929B (en) | Multi-stage multi-topology ship traffic complexity measurement method and system | |
| Sun et al. | Detecting Crime Types Using Classification Algorithms. | |
| Mostofi et al. | Explainable safety risk management in construction with unsupervised learning | |
| CN116910526A (en) | Model training method, device, communication equipment and readable storage medium | |
| CN114860953A (en) | Interpretable method based on few-sample relation prediction model | |
| CN115630732A (en) | City operation-oriented enterprise migration big data monitoring and early warning method and device | |
| CN113887115A (en) | Method, device, equipment and storage medium for safety detection of running state | |
| Fu et al. | Prediction of fishing vessel operation mode based on Stacking model fusion | |
| Zhu et al. | Marine environment monitoring based on virtual reality and fuzzy C-means clustering algorithm | |
| Chen et al. | Application of deep learning model in computer data mining intrusion detection |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210430 |
|
| WW01 | Invention patent application withdrawn after publication |