CN-121997186-A - Mine user portrait and classification method based on safety big data

CN121997186ACN 121997186 ACN121997186 ACN 121997186ACN-121997186-A

Abstract

The invention relates to a mine user portrait and classification method based on safety big data, and belongs to the technical field of mine safety and data processing. The method comprises the steps of establishing a mine user portrait basic data set, constructing a user portrait label and classifying safety risks. According to the method, through acquisition, integration and analysis of multi-dimensional data of the mine, comprehensive and accurate mine user images are constructed, scientific classification is carried out, so that powerful support is provided for mine safety management, safety risks are effectively identified, prevented and controlled, and the mine safety production level is improved.

Inventors

XU YANMEI
XU BIN
JIA PENGJIE
ZHANG PENG
WANG HUA
WANG KAICHAO

Assignees

宝信软件(云南)有限公司

Dates

Publication Date: 20260508
Application Date: 20251128

Claims (8)

1. A mine user portrait and classification method based on safety big data is characterized by comprising the following steps: s1, establishing a mine user portrait basic data set, namely acquiring effective characteristic data by acquiring and preprocessing mine multi-source data, wherein the multi-source data comprises miners data, equipment data, environment data and text data; s2, constructing a user portrait tag system: Based on the mine user portrait basic data set obtained in the step S1, a feature weight optimization method integrating dynamic weighting and cluster analysis is adopted to generate user portrait labels covering miner behavior features, equipment running states, environmental safety indexes and text derivatives; and S3, security risk classification, namely quantifying the user portrait label into a feature vector, inputting the feature vector into a random forest classification model, and calculating a comprehensive risk value based on a hierarchical analysis method and a D-S evidence theory to finally obtain the security risk level of the mine user.
2. The mining user portrait and classifying method based on safety big data according to claim 1 is characterized in that in S1, the preprocessing includes data cleaning, missing value complementation, noise reduction, text data structuring and data association, and the data association is to associate operation records in miner data with operation tracks.
3. The mine user portraits and classification method based on safety big data according to claim 1, characterized in that in S2, the characteristic data is given weight by adopting an improved weight calculation method, concretely: Wherein, the The adjusted weight is the weight of the j-th data in the data set i; c is a weighting factor, N represents the total number of data records in the whole data set; And calculating a result for the data frequency after dynamic weighting adjustment, wherein the result is as follows: in the formula, Frequency of occurrence of feature data in critical security events; the frequency of the feature data in daily monitoring data; Frequency of occurrence of feature data in the history data; the key security event weight coefficient; the weight coefficient of the data is monitored daily; identifying a key feature data set S from all features based on the cluster analysis result, and for the features belonging to the set S, carrying out initial weight Readjusting, and calculating the final weight omega ij′ ': wherein, gamma is cluster analysis and then adjusts the weight coefficient, The specific formula of (2) is: wherein S is a key feature data set, Representing the ith feature data in the dataset.
4. The mine user portrayal and classification method based on safety big data as set forth in claim 1, characterized in that in S2: the generation of the miner behavior characteristic labels comprises the steps of clustering the miner operation types by adopting a K-means algorithm and/or mining dangerous behavior sequences of the miners by adopting a PrefixSpan algorithm; The generation of the equipment running state label comprises the steps of detecting the abnormal state of the equipment health state by adopting an isolated forest algorithm and/or mining the association rule of the equipment fault risk by adopting an Apriori algorithm; The generation of the environmental safety index label comprises threshold judgment based on real-time parameters and/or trend early warning by adopting an ARIMA time sequence model; The generation of the text derived labels comprises the steps of calculating text keyword weights by adopting an improved TF-IWF algorithm, identifying text security topics and strengthening features by adopting an LDA topic model, and finally generating the labels based on weighted sequencing results.
5. The mine user portrayal and classification method based on safety big data as claimed in claim 1, wherein the specific method of S3 is as follows: quantizing the user portrait tag into a feature vector; learning an initial mapping relation from the tag feature vector to the security risk level by utilizing a random forest algorithm; Determining weights of four dimensions of miner behaviors, equipment states, environmental indexes and text information by adopting an analytic hierarchy process; The four dimensions are used as four independent evidence sources, and the Dempster combination rule of the D-S evidence theory is utilized to fuse the risk level probability distribution output by each evidence source; and calculating a comprehensive risk value according to the fused probability distribution, and judging a final security risk level according to a preset threshold value.
6. The mine user portrait and classifying method based on security big data as claimed in claim 1, further comprising S4 of dynamically updating user portrait tags and random forest models to adapt to dynamic changes of mine security state.
7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor performs the steps of the mine user representation and classification method based on security big data as claimed in claim 1.
8. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements the steps of the mine user representation and classification method based on security big data of claim 1.

Description

Mine user portrait and classification method based on safety big data Technical Field The invention belongs to the technical field of mine safety and data processing, and particularly relates to a mine user portrait and classification method based on safety big data. Background Along with the continuous expansion of mine exploitation scale and the promotion of intelligent mine construction, mine safety production management faces a plurality of challenges. The mine environment is complex and changeable, the miner operation behaviors are various, the running state of the equipment is changed in real time, and the situation of each element of the mine is comprehensively and accurately mastered, so that accurate safety management is realized, and the problem to be solved urgently in the current mine industry is solved. The existing mine management method has the defects in the aspects of data integration and analysis, and is difficult to effectively process and comprehensively analyze multidimensional information such as miners, equipment, environments and the like. The user portrait with comprehensive and accurate construction cannot be constructed, so that the recognition and prevention and control capabilities of mine safety risks are limited, and the requirements of modern mine safety management cannot be met. Therefore, a new method is needed to construct mine user images by using safety big data and classify the mine user images so as to improve the mine safety management level. Disclosure of Invention The invention aims to solve the defects of the prior art and provides a mine user portrait and classification method based on safety big data. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: A mine user portrait and classification method based on safety big data comprises the following steps: s1, establishing a mine user portrait basic data set, namely acquiring effective characteristic data by acquiring and preprocessing mine multi-source data, wherein the multi-source data comprises miners data, equipment data, environment data and text data; s2, constructing a user portrait tag system: Based on the mine user portrait basic data set obtained in the step S1, a feature weight optimization method integrating dynamic weighting and cluster analysis is adopted to generate user portrait labels covering miner behavior features, equipment running states, environmental safety indexes and text derivatives; and S3, security risk classification, namely quantifying the user portrait label into a feature vector, inputting the feature vector into a random forest classification model, and calculating a comprehensive risk value based on a hierarchical analysis method and a D-S evidence theory to finally obtain the security risk level of the mine user. Further, in S1, the preprocessing preferably includes data cleansing, missing value complementation, noise reduction, text data structuring and data association, and the data association is to associate an operation record in the miner data with a job track. Further, in S2, it is preferable that the feature data is weighted by an improved weight calculation method, specifically: Wherein, the The adjusted weight is the weight of the j-th data in the data set i; c is a weighting factor, N represents the total number of data records in the whole data set; And calculating a result for the data frequency after dynamic weighting adjustment, wherein the result is as follows: in the formula, Frequency of occurrence of feature data in critical security events; the frequency of the feature data in daily monitoring data; Frequency of occurrence of feature data in the history data; the key security event weight coefficient; the weight coefficient of the data is monitored daily; identifying a key feature data set S from all features based on the cluster analysis result, and for the features belonging to the set S, carrying out initial weight Readjusting, and calculating the final weight omega ij′': wherein, gamma is cluster analysis and then adjusts the weight coefficient, The specific formula of (2) is: wherein S is a key feature data set, Representing the ith feature data in the dataset. Further, it is preferable that in S2: the generation of the miner behavior characteristic labels comprises the steps of clustering the miner operation types by adopting a K-means algorithm and/or mining dangerous behavior sequences of the miners by adopting a PrefixSpan algorithm; The generation of the equipment running state label comprises the steps of detecting the abnormal state of the equipment health state by adopting an isolated forest algorithm and/or mining the association rule of the equipment fault risk by adopting an Apriori algorithm; The generation of the environmental safety index label comprises threshold judgment based on real-time parameters and/or trend early warning by adopting an ARIMA time sequence model; The generation of t