CN111860598B - Data analysis method and electronic equipment for identifying sports behaviors and relationships - Google Patents
Data analysis method and electronic equipment for identifying sports behaviors and relationships Download PDFInfo
- Publication number
- CN111860598B CN111860598B CN202010562538.7A CN202010562538A CN111860598B CN 111860598 B CN111860598 B CN 111860598B CN 202010562538 A CN202010562538 A CN 202010562538A CN 111860598 B CN111860598 B CN 111860598B
- Authority
- CN
- China
- Prior art keywords
- motion
- data
- model
- time sequence
- data set
- 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
- 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
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a data analysis method and electronic equipment for identifying motion behaviors and relationships, which comprises the following steps of firstly, collecting motion data through a plurality of sensors arranged at a mobile terminal to form an initial data set; secondly, preprocessing the initial data set based on a time series analysis method to obtain a time series characteristic data set; secondly, analyzing the time sequence characteristic data set to obtain fine characteristics for model construction; secondly, establishing a motion identity recognition model and a motion mode analysis model based on the extracted fine features; finally, reasonably and mathematically analyzing and explaining the characteristics in the motion identity recognition model to obtain a group of motion characteristics related to the motion style; and analyzing each user by combining the motion characteristics, constructing a group of user characteristics corresponding to the users, judging whether cheating behaviors such as running instead exist among the users by combining the motion identity prediction model and the motion characteristics of the users, and judging whether the users use the vehicles when running by using the motion mode prediction model.
Description
Technical Field
The invention relates to the technical field of data analysis, in particular to a data analysis method and electronic equipment for identifying sports behaviors and relations.
Background
The artificial intelligence community considers machine learning to be one of the branches in the field of artificial intelligence that can most represent intelligence. It is endeavoured to investigate how to improve the performance of the system itself by means of calculations, using experience. Machine learning is primarily concerned with algorithms on computers that generate models from data. By providing it with empirical data, it is able to generate models from these data. When a new situation occurs, the model can be provided for a corresponding judgment.
Random forests are a common class of machine learning algorithms, and are classifiers comprising a plurality of decision trees. It has the advantages that: the method is suitable for the problem of coupling multidimensional variables, and can process high-dimensional data and multiple co-linear data; the tolerance to abnormal values and noise is high, the model is not easy to over-fit, and higher simulation precision can be obtained; it can evaluate the importance of variables, etc., in determining categories.
Today, the relevant theories and techniques are perfected. Under the support of sufficient data sets, various classification problems can be well solved by utilizing a random forest method of machine learning.
However, many colleges and universities supervise and urge students to do physical exercises outside class through mobile phones APP such as 'running on a sidewalk' and 'sports world campus', but the algorithms and models provided by various software for identifying the movement behavior patterns of the students lack effectiveness, so that cheating situations such as running and riding occur in large quantity. At present, research on cheating on sports by utilizing a mobile phone APP is less, analysis for identifying sports behavior patterns by utilizing multiple sensors is less at home and abroad at present, most of the analysis is only stopped on a few sensors, and an effective method or result is not provided for prediction analysis of the multiple sensors.
Disclosure of Invention
The invention aims to solve the technical problem that the analysis of a motion behavior pattern recognized by using multiple sensors is less, most of the motion behavior patterns are only stopped on a few sensors, and an accurate analysis result cannot be obtained in the prior art, and provides a data analysis method and electronic equipment for motion behavior and relationship recognition.
The technical scheme adopted by the invention for solving the technical problem is as follows: constructing a data analysis method for athletic performance recognition, comprising the steps of:
s1, acquiring an initial data set for reflecting the motion behavior of a user;
s2, preprocessing each item of data in the initial data set based on a time sequence analysis method, extracting time sequence characteristics of each item of data, and constructing a first time sequence characteristic data set;
s3, based on the first time sequence feature data set, after data classification and initial motion mode analysis model construction are carried out by utilizing a machine learning classification algorithm, the influence degree of each classification feature on the accuracy rate of the current modeling model is analyzed, and the first time sequence feature data set is screened out according to the influence degree; then, constructing a motion pattern analysis model and a motion identity recognition model based on a second time sequence characteristic data set obtained by current screening;
wherein, screening the second time series characteristic data set specifically comprises: analyzing the weight value alpha of each time sequence feature in the first time sequence feature data set in the initial motion mode analysis model according to the influence degree of each classification feature on the accuracy rate of the current modeling model and the influence degree i I = 1.., M is the total number of timing characteristics;
weighting value alpha of each time sequence characteristic i Comparing with a preset weight threshold value beta, and selecting a weight value alpha i Time series characteristics greater than beta, forming a second time series characteristic data set;
wherein the weight value is alpha i The confirmation method specifically comprises the following steps: analyzing the weight alpha of each time sequence characteristic in the initial motion mode analysis model by adopting an average precision reduction method i ：
Adding random noise to each time sequence characteristic;
measuring the influence of the time sequence characteristics before and after random noise is added on the accuracy rate of the initial motion mode analysis model, and determining the weight value of each time sequence characteristic in the model according to the influence degree;
and S4, constructing a second time sequence characteristic data set of the data of the motion behaviors to be recognized by using the methods of the steps S2 and S3, inputting the second time sequence characteristic data set of the data of the motion behaviors to be recognized into the motion pattern analysis model and the motion identity recognition model constructed in the step S3, then recognizing the user identities based on the motion identity recognition model, and recognizing the corresponding motion behaviors of the user based on the motion pattern analysis model.
Another technical solution adopted to solve the technical problems of the present invention is: constructing a data analysis method for athletic performance recognition, which comprises the steps of arranging a plurality of sensors for monitoring athletic performance data at a mobile terminal;
and after the data monitored by the sensor is transmitted to the server through a wired or wireless network transmission protocol, executing the data analysis method for the motion behavior identification to realize the identification of the motion behavior of the corresponding user.
The invention discloses a data analysis method for realizing motion relation recognition according to the data analysis method for motion behavior recognition, which specifically comprises the following steps after step S4:
s5, performing mathematical analysis on the classification features in the second time sequence feature data set for constructing the motion identity recognition model, and explaining the classification features into motion features related to the motion style by combining with an actual motion scene;
and S6, identifying the motion relation among different users by utilizing the motion identity prediction model and combining a plurality of motion characteristics obtained based on the step S5.
The invention discloses an electronic device, comprising a memory and a processor, wherein:
the memory is used for storing computer-executable instructions corresponding to any one or more of the data analysis method for athletic performance recognition, the data analysis method for athletic performance recognition and the data analysis method for athletic relationship recognition;
the processor is configured to invoke and/or execute the computer-executable instructions stored in the memory.
In the data analysis method and the electronic equipment for identifying the sports behaviors and the relationships, a sports mode analysis model and a sports identity identification model are constructed, the identity and the sports mode of a sporter are subjected to predictive analysis, and the sports style of the sporter is reasonably explained in the form of user characteristics.
The implementation of the data analysis method and the electronic equipment for identifying the sports behaviors and the relationships has the following beneficial effects that:
the exercise behaviors of the user are analyzed in real time from a macroscopic angle, the possible exercise mode of the exerciser is effectively predicted, whether cheating behaviors such as running and riding exist or not is judged, and decision support of user exercise characteristic analysis and exercise mode identification can be provided for intelligent health hardware companies.
A method for analyzing the motion behavior pattern of an agent by using multiple sensors is provided, and higher precision is obtained in detection.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a data analysis method for athletic performance, relationship identification, in accordance with the present invention;
FIG. 2 is a flow chart of the model building process of the present invention;
fig. 3 is a structural diagram of an electronic device disclosed in the present invention.
Detailed Description
For a more clear understanding of the technical features, objects, and effects of the present invention, embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
Example 1:
in order to analyze the exercise behavior of the exerciser in time by combining with the application software, please refer to fig. 1, which is a flowchart of a data analysis method for identifying the exercise behavior and the relationship according to the present invention, and when analyzing the exercise behavior, by combining with fig. 1, the method specifically includes the following steps:
s1, acquiring an initial data set for reflecting the motion behavior of a user; the initial data set can be previously monitored by related data monitoring equipment and stored in mobile or fixed storage equipment, and when the initial data set needs to be acquired, the initial data set only needs to be copied from the corresponding storage equipment; the initial data set may also be transmitted to the processing terminal by means of an instant transmission, for example, an instant transmission of monitoring data by the sensor, and the processing terminal executes the next step. Each data of the initial data set consists of a feature set and a label, wherein the feature set is a data feature set formed after the copied data time sequence features are extracted; the label is divided into two cases, in the motion mode analysis model, the label refers to a motion mode, and in the motion identity recognition model, the label refers to the unique identity of the user.
S2, preprocessing each item of data in the initial data set based on a time sequence analysis method, extracting time sequence characteristics of each item of data, and constructing a first time sequence characteristic data set;
in the present step, the time series analysis method is a statistical method for dynamic data processing, and specifically performs symbolization processing on the obtained data, and performs time series feature extraction on the obtained symbolized time series data, where the extracted symbolized time series data includes:
the average, the median, the standard deviation, the normalized standard deviation, the range, the kurtosis, the skewness, the burstiness statistics, the coefficient of variation, the highlowmu statistics, the negative comparative number, the linear fit Root Mean Square Error (RMSE), the quadratic fit RMSE, the cubic fit RMSE, the average after 10% of the maximum and minimum values are removed, the customized skewness measure, the distance from the starting point to the first peak, the area from the starting point to the first peak, the curve length from the starting point to the first peak, the distance from the starting point to the first valley after the peak, the area from the starting point to the first valley after the peak, the curve length from the starting point to the first valley after the peak, the distance from the starting point to the maximum peak, the area from the starting point to the maximum peak, the curve length from the starting point to the maximum valley after the peak, the shannon entropy, and the approximate entropy are 30 features.
S3, based on the first time sequence feature data set, after data classification and initial motion mode analysis model construction are carried out by utilizing a machine learning classification algorithm, the influence degree of each classification feature on the accuracy rate of the current modeling model is analyzed, and the first time sequence feature data set is screened out according to the influence degree; then, constructing a motion pattern analysis model and a motion identity recognition model based on a second time sequence characteristic data set obtained by current screening;
in the current step, when an initial motion pattern analysis model, a motion pattern analysis model and a motion identity recognition model are constructed, a random forest algorithm is considered preferentially, the random forest algorithm is a classifier which trains and predicts samples by using a plurality of trees, and the output category of the random forest algorithm is determined by the mode of the category output by individual trees.
In this embodiment, the random forest algorithm is considered as follows:
the classifier has high accuracy, and can process a large number of input variables (the input variables are the first time-series feature data set in this embodiment); and, it can evaluate the importance of the input variables when determining the category; and, for an unbalanced classification data set, it can balance errors.
When establishing a motion identity recognition model and a motion pattern analysis model based on the extracted fine features for performing fine division on the initial motion pattern analysis model at present, consider:
s31, analyzing the weight value alpha of each time sequence feature in the first time sequence feature data set in the initial motion mode analysis model according to the influence degree of each classification feature on the accuracy rate of the initial motion mode and the influence degree i I = 1.., M is the total number of timing characteristics; specifically, the method comprises the following steps:
in this embodiment, an average precision reduction method is used to analyze the weight α occupied by each time sequence feature in the initial motion pattern analysis model i Wherein, the execution process of the average precision reduction method is as follows: 1. adding random noise to each time sequence characteristic; 2. and measuring the influence of the time sequence characteristics before and after the random noise is added on the accuracy rate of the initial motion mode analysis model, and determining the weight value of each time sequence characteristic in the model according to the influence degree.
In the specific implementation, it is obvious to those skilled in the art that:
for unimportant features, the model accuracy is not affected much after noise is added, but for important features, the accuracy is significantly affected. Based on this principle, the weight value of each feature in the model can be obtained.
In this embodiment, it may also be considered that, when the weight of each feature cannot be determined based on the model precision, an average precision reduction method may be used to disorder the data sequence in the corresponding data set, or some error terms may be added to make the modified data have a more obvious change than the data before modification. If the importance of a certain characteristic is high, the model precision is obviously reduced by disturbing the original data sequence; on the contrary, if the importance of a certain feature is small, the precision of the model before and after the data sequence is disturbed does not change greatly.
S32, weighting value alpha of each time sequence characteristic i Comparing with a preset weight threshold value beta, and selecting a weight value alpha i The time series characteristics larger than beta constitute a second time series characteristic data set. Wherein: in order to improve the execution efficiency of the algorithm, all feature weight values can be preferentially sorted according to the size before the weight value comparison is carried out, and the feature data can be directly selected after the weight threshold value beta is determined.
And S4, constructing a second time sequence characteristic data set of the data of the motion behaviors to be recognized by using the methods in the steps S2 and S3, inputting the second time sequence characteristic data set of the data of the motion behaviors to be recognized into the motion pattern analysis model and the motion identity recognition model constructed in the step S3, then performing user identity recognition based on the motion identity recognition model, and performing corresponding user motion behavior recognition based on the motion pattern analysis model.
At present, experiments can be carried out by using the exercise data of 20 runners, the unique identity of each runner is used as a label corresponding to the runner, the label is an integer between 0 and 19, and after the division is carried out according to the division basis, one runner corresponds to and has a unique label.
In the present step, when analyzing and predicting the input data of the exercise behavior to be recognized by using the exercise pattern analysis model, the probabilities of the plurality of exercise patterns may be obtained. The model can analyze whether the user has cheating behaviors such as riding and the like while judging the motion mode of the user.
In the embodiment, an analysis model for predicting the human motor behavior pattern is constructed by using a random forest method in time series symbolization extraction and machine learning, and the motor behavior pattern of each individual is analyzed by taking the individual as a basic scale, so that the blank in the prior art is made up.
Example 2:
referring to fig. 1, when the data analysis method for athletic performance identification described in embodiment 1 is implemented to identify a user athletic relationship, specifically, after step S4, the method further includes:
s5, analyzing the classification features of the second time sequence feature data set for constructing a motion identity recognition model, and explaining the classification features into motion features related to motion styles by combining with an actual motion scene;
and S6, identifying the motion relation among different users by utilizing the motion identity prediction model and combining a plurality of motion characteristics obtained based on the step S5.
As another implementation manner in this embodiment, when step S5 is executed, a user feature model may be constructed by using the obtained motion features; and identifying the motion relation among different users by combining the motion identity prediction model and the motion characteristics of the users analyzed by the user characteristic model.
In the present embodiment, reasonable mathematical analysis is performed on the features in the motion identification model, and the analyzed features are considered to be interpreted as a group of motion features related to the motion (e.g., running) style; and after the relation between each user is analyzed by combining the motion characteristics, constructing a user characteristic model. By utilizing the motion identity prediction model and combining the user motion characteristics analyzed by the user characteristic model, whether cheating behaviors such as race generation and the like exist among users can be effectively judged.
In the embodiment, the user characteristic model is provided for the first time, so that the running style of the user can be intuitively analyzed; the motion relation between the users can be effectively judged by combining the motion identity recognition model and the user characteristic model, and the motion relation can be expressed as whether a running generation phenomenon exists between the two users.
Example 3:
referring to fig. 1, when the data analysis method for identifying an exercise behavior and a relationship described in embodiments 1 and 2 is used to implement user exercise behavior identification, the method includes:
a1, arranging a plurality of sensors for monitoring motion behavior data at a mobile end; the mobile terminal can be a mobile phone, an intelligent bracelet and other portable mobile devices;
a2, after the data monitored by the sensor is transmitted to a server through a wired or wireless network transmission protocol (for example, tcp protocol), the data analysis method for the exercise behavior recognition as described in embodiments 1 and 2 is executed, so as to realize recognition of the exercise behavior of the corresponding user.
In the embodiment, at home and abroad, few analyses for identifying the motion behavior pattern by using multiple sensors are considered, most of the analyses stay on only a few sensors, and a method or result for predictive analysis of the multiple sensors is not available.
The mobile terminal is provided with a plurality of sensors of different types, wherein the sensors comprise an acceleration sensor, a direction sensor, a gyroscope sensor, a magnetic field sensor, a gravity sensor and a linear acceleration sensor.
Based on embodiment 1, when an initial data set is constructed, a corresponding storage field is set in the initial data set, and data transmitted by a sensor is directly stored in the storage field, so that the construction of the data set can be completed.
The athletic performance data monitored based on the sensors includes:
three-axis components of an acceleration sensor, three-axis components of a linear acceleration sensor, three-axis components of a direction sensor, three-axis components of a gravity sensor, three-axis components of a magnetic field sensor and three-axis components of a gyroscope sensor; there are 18 items of data.
Based on embodiment 1, when 18 collected items of data are symbolized by using a time series symbolization algorithm, the mean value, the median value, the standard deviation, the normalized standard deviation, the range, the kurtosis, the skewness, the burstiness statistics, the coefficient of variation, the highlowmu statistics, the negative comparative likelihood, the linear fitting Root Mean Square Error (RMSE), the quadratic fitting RMSE, the cubic fitting RMSE, the mean value after 10% of the maximum minimum value is removed, the customized skewness measurement, the distance from the starting point to the first peak, the area from the starting point to the first peak, the curve length from the starting point to the first peak, the distance from the starting point to the first valley after the peak, the area from the starting point to the first valley after the peak, the curve length from the starting point to the maximum peak, the distance from the starting point to the maximum valley after the peak, the area from the starting point to the maximum valley after the peak, the starting point to the maximum valley after the peak, the maximum entropy, the entropy, and the entropy are all 30 characteristics in total.
Currently, after feature extraction, the data dimension of the initial data set becomes 540-dimensional high-dimensional time series feature data. After feature selection in the motion identification model based on the embodiment 1-2, a 10-dimensional data set that can be put into the model for prediction can be obtained.
At present, it can be reflected that when the present embodiment is used for processing mass data, effective data are screened out from mass data in a centralized manner after a random forest classification algorithm and an average precision reduction method are combined, and the execution efficiency of the algorithm is further improved.
Example 4:
the system structure of an electronic device disclosed in this embodiment may refer to fig. 3, where the device includes a memory and a processor, where:
the memory is used for storing computer-executable instructions corresponding to any one or more of the data analysis method for athletic performance recognition described in any one of the above, the data analysis method for athletic performance recognition described above, and the data analysis method for athletic relationship recognition described above;
the processor is configured to invoke and/or execute the computer-executable instructions stored in the memory.
Referring to fig. 2, a flow chart of a model building process according to the present invention will be further described with reference to embodiments 1-3, including:
firstly, collecting motion data of volunteers by using a mobile terminal;
secondly, extracting time sequence characteristics of the data by using a time sequence symbolization algorithm, and selecting m characteristics with the maximum weight (in combination with a specific implementation scene, 10 characteristics are selected for subsequent steps in the embodiment);
secondly, constructing a random forest model; further refining the random forest model based on the random forest model to obtain a motion mode identification model and a motion identity identification model;
finally, when the movement pattern is identified, calculating the probability of each movement pattern based on the movement pattern identification model, and judging whether cheating behaviors such as riding exist or not;
finally, when the motion identity is identified, the identity of the user is predicted according to the motion identity identification model, and cheating phenomena such as generation running and the like can exist for further determination; considering a reasonable mathematical analysis of features in an athletic identification model, considering the interpretation of these analyzed features into a set of athletic features related to an athletic (e.g., running) style; and constructing a user feature model after analyzing the relation between each user by combining the motion features. By utilizing the motion identity prediction model and combining the user motion characteristics analyzed by the user characteristic model, whether cheating behaviors such as race generation and the like exist among users can be effectively judged.
The data analysis method and the electronic equipment for identifying the motion behaviors and the relationships disclosed by the invention can be used for carrying out instant analysis on the motion behaviors of the user from a macroscopic view, effectively predicting the possible motion mode of the sporter, judging whether cheating behaviors such as running and riding exist or not, and providing decision support for motion characteristic analysis and motion mode identification of the user for an intelligent health hardware company. Furthermore, a method for analyzing the behavioral patterns of the human locomotion by using multiple sensors is proposed, and high accuracy is obtained in detection.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (7)
1. A data analysis method for athletic performance recognition, comprising the steps of:
s1, acquiring an initial data set for reflecting the motion behavior of a user;
s2, preprocessing each item of data in the initial data set based on a time sequence analysis method, extracting time sequence characteristics of each item of data, and constructing a first time sequence characteristic data set;
s3, based on the first time sequence feature data set, after data classification and initial motion mode analysis model construction are carried out by utilizing a machine learning classification algorithm, the influence degree of each classification feature on the accuracy rate of the current modeling model is analyzed, and the first time sequence feature data set is screened out according to the influence degree; then, constructing a motion pattern analysis model and a motion identity recognition model based on a second time sequence characteristic data set obtained by current screening;
wherein, screening the second time series characteristic data set specifically comprises: analyzing the weight value alpha of each time sequence feature in the first time sequence feature data set in the initial motion mode analysis model according to the influence degree of each classification feature on the accuracy rate of the current modeling model and the influence degree i I = 1.., M is the total number of timing characteristics;
weighting value alpha of each time sequence characteristic i Comparing with a preset weight threshold value beta, and selecting a weight value alpha i Time series characteristics greater than beta, forming a second time series characteristic data set;
wherein the weight value alpha i The confirmation method specifically comprises the following steps: analyzing the weight alpha occupied by each time sequence characteristic in the initial motion mode analysis model by adopting an average precision reduction method i ：
Adding random noise to each time sequence characteristic;
measuring the influence of the time sequence characteristics before and after random noise is added on the accuracy rate of the initial motion mode analysis model, and determining the weight value of each time sequence characteristic in the model according to the influence degree;
and S4, constructing a second time sequence characteristic data set of the data of the motion behaviors to be recognized by using the methods of the steps S2 and S3, inputting the second time sequence characteristic data set of the data of the motion behaviors to be recognized into the motion pattern analysis model and the motion identity recognition model constructed in the step S3, then recognizing the user identities based on the motion identity recognition model, and recognizing the corresponding motion behaviors of the user based on the motion pattern analysis model.
2. The data analysis method for identifying an athletic performance of claim 1, wherein in step S2, the time-series analysis method is to perform a symbolization process on the acquired data and extract a time-series feature for the obtained symbolized time-series data.
3. The data analysis method for athletic performance recognition according to claim 1, wherein in step S3, a random forest algorithm is used to model each item of data in the first and second time series characteristic data sets to obtain an initial motion pattern analysis model and a motion identity recognition model.
4. A data analysis method for athletic performance recognition according to any one of claims 1-3, wherein after step S4, the method further comprises:
s5, analyzing the classification features of the second time sequence feature data set for constructing a motion identity recognition model, and explaining the classification features into motion features related to motion styles by combining with an actual motion scene;
and S6, identifying the motion relation among different users by utilizing the motion identity prediction model and combining a plurality of motion characteristics obtained based on the step S5.
5. A data analysis method for motion relation recognition comprises the steps that a plurality of sensors used for monitoring motion behavior data are arranged at a mobile end; the sensors arranged at the moving end comprise an acceleration sensor, a direction sensor, a gyroscope sensor, a magnetic field sensor, a gravity sensor and a linear acceleration sensor;
the athletic performance data monitored based on the sensors includes:
three-axis components of an acceleration sensor, three-axis components of a linear acceleration sensor, three-axis components of a direction sensor, three-axis components of a gravity sensor, three-axis components of a magnetic field sensor and three-axis components of a gyroscope sensor;
transmitting the motion behavior data to a server by using a tcp (Transmission control protocol);
the method is characterized in that after data monitored by a sensor is transmitted to a server through a wired or wireless network transmission protocol, the data analysis method for the motion behavior recognition according to any one of claims 1 to 3 is executed, and recognition of the motion relation of a corresponding user is achieved.
6. The data analysis method for motion relationship recognition according to claim 5, wherein a user feature model is constructed using the obtained motion features; and identifying the motion relation among different users by combining the motion identity prediction model and the motion characteristics of the users analyzed by the user characteristic model.
7. An electronic device comprising a memory and a processor, characterized in that:
the memory is used for storing computer executable instructions corresponding to any one of the data analysis method for the athletic performance recognition according to any one of claims 1-4 and the data analysis method for the athletic relationship recognition according to claim 5;
the processor is configured to invoke and/or execute the computer-executable instructions stored in the memory.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010562538.7A CN111860598B (en) | 2020-06-18 | 2020-06-18 | Data analysis method and electronic equipment for identifying sports behaviors and relationships |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010562538.7A CN111860598B (en) | 2020-06-18 | 2020-06-18 | Data analysis method and electronic equipment for identifying sports behaviors and relationships |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111860598A CN111860598A (en) | 2020-10-30 |
| CN111860598B true CN111860598B (en) | 2023-02-28 |
Family
ID=72986919
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010562538.7A Active CN111860598B (en) | 2020-06-18 | 2020-06-18 | Data analysis method and electronic equipment for identifying sports behaviors and relationships |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111860598B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112883348A (en) * | 2021-02-05 | 2021-06-01 | 西安电子科技大学 | Identity verification method, device and equipment based on motion characteristics and computer storage medium |
| CN113128131A (en) * | 2021-05-17 | 2021-07-16 | 中国银行股份有限公司 | Step-counting cheating detection method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101976451A (en) * | 2010-11-03 | 2011-02-16 | 北京航空航天大学 | Motion control and animation generation method based on acceleration transducer |
| CN103377242A (en) * | 2012-04-25 | 2013-10-30 | Tcl集团股份有限公司 | User behavior analysis method, user behavior analytical prediction method and television program push system |
| CN107016233A (en) * | 2017-03-14 | 2017-08-04 | 中国科学院计算技术研究所 | The association analysis method and system of motor behavior and cognitive ability |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9050528B2 (en) * | 2006-07-14 | 2015-06-09 | Ailive Inc. | Systems and methods for utilizing personalized motion control in virtual environment |
| CN104299007A (en) * | 2014-09-17 | 2015-01-21 | 电子科技大学 | Classifier training method for behavior recognition |
| US9934443B2 (en) * | 2015-03-31 | 2018-04-03 | Daon Holdings Limited | Methods and systems for detecting head motion during an authentication transaction |
| CN104881655B (en) * | 2015-06-03 | 2018-08-28 | 东南大学 | A kind of human behavior recognition methods based on the fusion of multiple features time-space relationship |
-
2020
- 2020-06-18 CN CN202010562538.7A patent/CN111860598B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101976451A (en) * | 2010-11-03 | 2011-02-16 | 北京航空航天大学 | Motion control and animation generation method based on acceleration transducer |
| CN103377242A (en) * | 2012-04-25 | 2013-10-30 | Tcl集团股份有限公司 | User behavior analysis method, user behavior analytical prediction method and television program push system |
| CN107016233A (en) * | 2017-03-14 | 2017-08-04 | 中国科学院计算技术研究所 | The association analysis method and system of motor behavior and cognitive ability |
Non-Patent Citations (3)
| Title |
|---|
| 基于加速度传感器的人体运动状态识别研究;第03期;《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》;20160315;第7-30、38-43页 * |
| 第03期.基于加速度传感器的人体运动状态识别研究.《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》.2016, * |
| 运动识别中基于主题的特征构建方法;郭浩东等;《浙江大学学报(工学版)》;20160630(第06期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111860598A (en) | 2020-10-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110443002A (en) | A kind of Deformation of Steep Slopes prediction technique and system | |
| CN109325691A (en) | Abnormal behaviour analysis method, electronic equipment and computer program product | |
| CN101650687B (en) | Large-scale parallel program property-predication realizing method | |
| CN111860598B (en) | Data analysis method and electronic equipment for identifying sports behaviors and relationships | |
| CN111310918B (en) | Data processing method, device, computer equipment and storage medium | |
| CN112001110A (en) | Structural damage identification monitoring method based on vibration signal space real-time recursive graph convolutional neural network | |
| CN110647995A (en) | Rule training method, device, equipment and storage medium | |
| CN108304853A (en) | Acquisition methods, device, storage medium and the electronic device for the degree of correlation of playing | |
| CN110138766A (en) | Longicorn must be in conjunction with the network inbreak detection method of random forest | |
| CN111984544B (en) | Device performance test method and device, electronic device and storage medium | |
| CN108717548B (en) | Behavior recognition model updating method and system for dynamic increase of sensors | |
| CN109117352B (en) | Server performance prediction method and device | |
| CN110084291B (en) | Student behavior analysis method and device based on big data extreme learning | |
| CN115795329A (en) | Power utilization abnormal behavior analysis method and device based on big data grid | |
| CN109121133B (en) | Location privacy protection method and device | |
| CN106484913A (en) | Method and server that a kind of Target Photo determines | |
| CN112888008A (en) | Base station abnormity detection method, device, equipment and storage medium | |
| CN116680613A (en) | Human activity recognition comprehensive optimization method based on multi-scale metric learning | |
| CN116703466A (en) | System access quantity prediction method based on improved wolf algorithm and related equipment thereof | |
| CN115809795A (en) | Digitalized production team bearing capacity evaluation method and device | |
| CN110096708A (en) | A kind of determining method and device of calibration collection | |
| CN115348190A (en) | Internet of things equipment detection method, system and equipment | |
| CN115270983A (en) | Switch cabinet fault prediction method based on AdaBoost-RBF algorithm | |
| CN115169705A (en) | Distribution time length prediction method and device, storage medium and computer equipment | |
| CN115687034A (en) | Service system plane availability judgment method and device |
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 |