CN111967494B - Multi-source heterogeneous data analysis method for guard security of large movable public security system - Google Patents
Multi-source heterogeneous data analysis method for guard security of large movable public security system Download PDFInfo
- Publication number
- CN111967494B CN111967494B CN202010625497.1A CN202010625497A CN111967494B CN 111967494 B CN111967494 B CN 111967494B CN 202010625497 A CN202010625497 A CN 202010625497A CN 111967494 B CN111967494 B CN 111967494B
- Authority
- CN
- China
- Prior art keywords
- data
- word
- track
- text
- analysis
- 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
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000007405 data analysis Methods 0.000 title claims abstract description 14
- 238000004458 analytical method Methods 0.000 claims abstract description 25
- 230000008451 emotion Effects 0.000 claims abstract description 15
- 230000000694 effects Effects 0.000 claims abstract description 12
- 238000000605 extraction Methods 0.000 claims abstract description 9
- 230000035515 penetration Effects 0.000 claims abstract description 3
- 238000011160 research Methods 0.000 claims abstract description 3
- 239000013598 vector Substances 0.000 claims description 18
- 230000015654 memory Effects 0.000 claims description 13
- 238000012549 training Methods 0.000 claims description 13
- 238000012552 review Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000012216 screening Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 4
- 238000012544 monitoring process Methods 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 2
- 238000013145 classification model Methods 0.000 claims description 2
- 238000004140 cleaning Methods 0.000 claims description 2
- 230000002996 emotional effect Effects 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 claims description 2
- 230000002265 prevention Effects 0.000 abstract description 3
- 238000013135 deep learning Methods 0.000 abstract description 2
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 231100000614 poison Toxicity 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000003440 toxic substance Substances 0.000 description 1
Classifications
-
- 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
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- 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
-
- 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
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
- G06Q50/265—Personal security, identity or safety
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Tourism & Hospitality (AREA)
- Probability & Statistics with Applications (AREA)
- Educational Administration (AREA)
- Databases & Information Systems (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Primary Health Care (AREA)
- Marketing (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Computer Security & Cryptography (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a multi-source heterogeneous data analysis method for guard security of a large-scale movable public security system, which is used for carrying out text emotion analysis, criminal trend personnel portraits, criminal trend personnel track analysis and text information extraction in a deep learning mode. Firstly, emotion analysis is carried out on the network social texts, accounts which release bad comments are screened out, and all contents released by the accounts and search records are analyzed to obtain information such as penetration, age, cultural degree and the like of account holders. Track analysis and text information extraction are carried out on key personnel who frequently come to and go from an activity holding place and a place where people tend to go home in the near future so as to effectively fight crimes, reduce the workload of first-line dry polices, effectively ensure the smooth holding of large-scale activities and provide a solution for further research of crime prevention methods of public security systems.
Description
Technical Field
The invention discloses a multi-source heterogeneous data analysis method for guard security of a large-scale movable public security system based on deep learning, which is mainly used for related works such as major movable security, case detection assistance and the like, and belongs to the field of public safety big data mining and analysis.
Background
With the improvement of international status of China, outsiders are frequent, and how to ensure security work in high-specification major activities has become a difficult problem for public security guard departments. At present, the large-scale activity safety guard work is still mainly based on the traditional sea tactics, and various emergency situations are passively defended. With the complexity of security situations and the diversification of criminal means, this mode has failed to meet the actual working needs. Firstly, the ever-perfecting chemical technology makes toxic substances difficult to detect, has high concealment, has high killing power and the like; secondly, terrorism tends to be increasingly rampant worldwide, and threatens the safety of people worldwide at any time; the security situation is complex. The advancement of public security work informatization in China causes a great amount of public security data to be accumulated, and the development of the Internet industry in China also accumulates a great amount of user data. Along with the coming of the artificial intelligence era, how to effectively enable data to serve the public security system becomes the trend of the public security work informatization construction future development. The method is mainly aimed at the construction and development of public security crime prevention systems. The application of the big data analysis technology in the security guard new mode is beneficial to the development of the hidden information of the data, and combines the factors of people, things, places, things, organizations and the like of the cases to establish an early warning model, thereby providing scientific basis for predicting and preventing crimes.
The multi-source heterogeneous data analysis method is built, the police data analysis model is built, criminals are killed in the cradle, first-line police workload is reduced, efficiency is improved, risks are reduced to the lowest possible, a solid rear shield is provided for smooth progress of high-specification important activities, scientific decision assistance is provided for a leading layer, an empirically-based mode is eliminated, scientific decisions are made based on data, misjudgment rate is reduced, and efficiency is improved. Effectively ensuring the smooth running of high-specification major activities.
Disclosure of Invention
The invention provides a multi-source heterogeneous data analysis method for crime prevention of a public security system, which is divided into a Chinese text emotion analysis model, a track analysis model and a criminal trend personnel portrait and Chinese information extraction model. The Chinese text emotion analysis model takes a natural language text as input of a long-short-time memory network, continuously optimizes a network weight coefficient through training and iteration of a neural network, and outputs the model as emotion judgment of the text, emotion score and crime type prejudgment. And carrying out first-round screening on the mass accounts. After the first round of screening, criminal trend personnel portraits are carried out on the residual accounts, the released contents of the locked accounts and the read searching records are input into a long-short-term memory network to carry out Chinese text multi-label classification, and personal information such as gender, age, native place, cultural degree and the like of the account holders are judged; after judging the basic characteristics of crime prone personnel, carrying out track analysis on abnormal track persons which frequently come and go to a large-scale activity to be held in the near future, clustering tracks by using a DBSCAN algorithm, carrying out data analysis by combining the time generated by the track points, and further screening target groups; after the key crowd is locked, the social account content is tracked and information is extracted, and information such as the time, place, character, telephone and the like of the draft is extracted from unstructured text data. The crime description related to the invention refers to a special technical crowd needing to be tracked, and the crowd which is in line with the technical theme of the invention and can influence the guard security of a large-scale movable public security system.
The invention is realized by the following technical scheme, which comprises the following steps:
step 1: carrying out emotion analysis on text content released by the network social account, and finding out an account with criminal tendency; performing emotion assessment on the network social text data by utilizing an LSTM algorithm under a Keras framework, and primarily screening criminal trend groups;
step 2: portraying criminal tendency personnel, performing Chinese text multi-label classification on the key account content and the key account search record screened in the step 1 by utilizing an LSTM algorithm under a Tensorflow framework, and deducing the academy, age, gender and native place penetration information of the primarily screened criminal tendency crowd;
step 3: a crime tendency personnel track analysis and information extraction model; performing track analysis on criminal inclined personnel by adopting a DBSCAN algorithm, extracting information from a social account by utilizing Python, clustering track points of people frequently going to and from a target place, and performing comprehensive analysis by combining track generation time; and monitoring the social account, programming by using a Python language, and extracting information from chat record information of the target account to obtain information.
The technical principle of the invention is as follows:
1. and vectorizing the text data, inputting the vectorized text data into an LSTM network for data processing, and outputting emotion analysis results or labels. The long-short time memory network plus Softmax classifier is an effective classification method, and after the existing classified data are repeatedly trained, the unknown data can be judged to be classified by using the training result through the characteristic value. Compared with other classification methods, the method has the advantages of good training effect, simple operation and suitability for Chinese text multi-label classification in performance and accuracy.
2. Clustering the track points by using a DBSCAN algorithm, and deducing the attribute of the clustered points by combining time. DBSCAN is a density space based clustering algorithm that does not require determining the number of clusters, but rather based on the number of data-speculative clusters, can generate clusters for arbitrary shapes. The greatest characteristic of DBSCAN is that the variety of clusters is not needed to be determined in advance, and the clusters and the outliers are found out through a density-based method. Most points in the class are analyzed, and outliers in the track are also analyzed, so that information omission is avoided to the greatest extent.
Drawings
FIG. 1 is a diagram of a Word2Vec neural network
FIG. 2 is a schematic diagram of a long and short term memory network
FIG. 3 is a diagram of a text emotion analysis model
Detailed Description
The following describes in detail examples of the present invention, which are implemented on the premise of the technical solution of the present invention, and give detailed embodiments and specific operation procedures, but the scope of protection of the present invention is not limited to the following examples.
Step 1: firstly, a Chinese text emotion analysis model is constructed, emotion analysis and possibly related crime type extraction are carried out on text data of a network social account, and massive social accounts are initially screened, wherein the method comprises the following steps:
(1) Word2Vec algorithm compresses data size while capturing context information; word2Vec is actually two different methods: continuous Bag of Words and Skip-gram; the goal of CBOW is to predict the current word from context; skip-gram predicts context based on current word; initially, each word is a random N-dimensional vector; after training, the Word2Vec algorithm obtains the optimal vector of each Word, namely a Word vector, by using a CBOW or Skip-gram method; the word vector has captured the context information; discovering the relation between words by using a basic algebraic formula; these word vectors replace the bag of words model to predict the emotional state of the unknown data;
(2) The LSTM network sends word vectors into the neural network, and the LSTM has two lines, one open line containing the current data stream; a dark line containing the memory flow of the cell itself; in the "input gate", the influence of the cell memory is controlled according to the current data flow; next, in the "forget gate", the memory and data flow of this cell is updated; then generating output updated memory and data stream in the output gate;
(3) Loading a training file and performing Chinese word segmentation; creating a word dictionary, and returning an index of each word, a word vector and a word index corresponding to each sentence; and realizing an LSTM network by adopting a keras library in Python and training the network to store.
Step 2: and (3) extracting text content and search records released by the crime tendency accounts screened in the step one, and portraying crime tendency personnel.
(1) The text class is converted to Id, facilitating the training of the classification model at a later time.
(2) After converting the text category into Id, because the data are Chinese, preprocessing the Chinese, and cleaning the data before using the text data;
(3) After the data preprocessing is completed, modeling work of the LSTM is started next: to carry out vectorization processing on the cut_review data, each cut_review is converted into a vector of an integer sequence, 50000 words which are most frequently used are set, and the maximum word number of each cut_review is set to be 250;
(4) Defining a sequence model of LSTM: the first layer of the model is an embedded layer, representing each word using a vector of length 100; the spatldropout 1D layer randomly sets the ratio of the input unit to 0 each time it is updated in training; the LSTM layer comprises 100 memory units, and the output layer is a full connection layer comprising 10 classifications; because of the Chinese text multi-label classification, the activation function is set to 'softmax' and the loss function is the classification cross entropy.
Step 3: clustering the track points by using a DBSCAN algorithm for criminal tendency personnel, and deducing the attribute of the clustered points by combining time. Structured content extraction of unstructured text is achieved using the Python language.
A crime tendency personnel track analysis and information extraction model; the personnel portrait is obtained after the processing in the last step, track analysis is carried out on personnel frequently going to and from the place where the activities are held and the place where the activities are carried out, the track data are selected from Geolife track data of Microsoft Asian research institute, the GPS track of the Geolife track data is represented by a series of time stamp points, and each time stamp point comprises latitude, longitude and altitude information; firstly, reading in data, selecting required longitude and latitude data, displaying a user track on a *** map, and then clustering a data set by using a DBSCAN algorithm and calculating a central point of each cluster; each cluster represents frequent visits by the user to the area; suppose that the user's work and residence are in the 4 clusters; re-reading the data, looking at the hours distribution in each cluster and showing the inferences of the work sites and residence sites on the graph; and monitoring the social account, and extracting information in the text.
Claims (4)
1. A multi-source heterogeneous data analysis method for guard security of a large-scale movable public security system is characterized by comprising the following steps of: the method comprises the following steps:
step 1: carrying out emotion analysis on text content released by the network social account, and finding out an account with criminal tendency; performing emotion assessment on the network social text data by utilizing an LSTM algorithm under a Keras framework, and primarily screening criminal trend groups;
step 2: portraying criminal tendency personnel, performing Chinese text multi-label classification on the key account content and the key account search record screened in the step 1 by utilizing an LSTM algorithm under a Tensorflow framework, and deducing the academy, age, gender and native place penetration information of the primarily screened criminal tendency crowd;
step 3: a crime tendency personnel track analysis and information extraction model; performing track analysis on criminal inclined personnel by adopting a DBSCAN algorithm, extracting information from a social account by utilizing Python, clustering track points of people frequently going to and from a target place, and performing comprehensive analysis by combining track generation time; and monitoring the social account, programming by using a Python language, and extracting information from chat record information of the target account to obtain information.
2. The multi-source heterogeneous data analysis method for guard security of a large-scale movable public security system according to claim 1, which is characterized in that: emotion analysis is carried out on text content released by a network social account, and an account with crime tendency is found out, wherein the concrete method comprises the following steps:
text vectorization: word2Vec algorithm compresses data size while capturing context information; word2Vec is actually two different methods: continuous Bag ofWords and Skip-gram; the goal of CBOW is to predict the current word from context; skip-gram predicts context based on current word; initially, each word is a random N-dimensional vector; after training, the Word2Vec algorithm obtains the optimal vector of each Word, namely a Word vector, by using a CBOW or Skip-gram method; the word vector has captured the context information; discovering the relation between words by using a basic algebraic formula; these word vectors replace the bag of words model to predict the emotional state of the unknown data;
the LSTM network sends word vectors into the neural network, and the LSTM has two lines, one open line containing the current data stream; a dark line containing the memory flow of the cell itself; in the "input gate", the influence of the cell memory is controlled according to the current data flow; next, in the "forget gate", the memory and data flow of this cell is updated; then generating output updated memory and data stream in the output gate;
the algorithm flow is as follows: loading a training file and performing Chinese word segmentation; creating a word dictionary, and returning an index of each word, a word vector and a word index corresponding to each sentence; and realizing an LSTM network by adopting a keras library in Python and training the network to store.
3. The multi-source heterogeneous data analysis method for guard security of a large-scale movable public security system according to claim 1, which is characterized in that: portraying criminal trend personnel, realizing accurate striking, effectively preventing, the method is as follows:
s1, converting the text category into Id, so as to facilitate the training of a later classification model;
s2, after converting the text category into Id, preprocessing the Chinese as the data are Chinese, and cleaning the data before using the text data;
after the S3 data preprocessing is completed, modeling work of LSTM is started next: to carry out vectorization processing on the cut_review data, each cut_review is converted into a vector of an integer sequence, 50000 words which are most frequently used are set, and the maximum word number of each cut_review is set to be 250;
s4 defines a sequence model of LSTM: the first layer of the model is an embedded layer, representing each word using a vector of length 100; the spatldropout 1D layer randomly sets the ratio of the input unit to 0 each time it is updated in training; the LSTM layer comprises 100 memory units, and the output layer is a full connection layer comprising 10 classifications; because of the Chinese text multi-label classification, the activation function is set to 'softmax' and the loss function is the classification cross entropy.
4. The multi-source heterogeneous data analysis method for guard security of a large-scale movable public security system according to claim 1, which is characterized in that: a crime tendency personnel track analysis and information extraction model; the personnel portrait is obtained after the processing in the last step, track analysis is carried out on personnel frequently going to and from the place where the activities are held and the place where the activities are carried out, the track data are selected from Geolife track data of Microsoft Asian research institute, the GPS track of the Geolife track data is represented by a series of time stamp points, and each time stamp point comprises latitude, longitude and altitude information; firstly, reading in data, selecting required longitude and latitude data, displaying a user track on a *** map, and then clustering a data set by using a DBSCAN algorithm and calculating a central point of each cluster; each cluster represents frequent visits by the user to the area; suppose that the user's work and residence are in the 4 clusters; re-reading the data, looking at the hours distribution in each cluster and showing the inferences of the work sites and residence sites on the graph; and monitoring the social account, and extracting information in the text.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010625497.1A CN111967494B (en) | 2020-07-01 | 2020-07-01 | Multi-source heterogeneous data analysis method for guard security of large movable public security system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010625497.1A CN111967494B (en) | 2020-07-01 | 2020-07-01 | Multi-source heterogeneous data analysis method for guard security of large movable public security system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111967494A CN111967494A (en) | 2020-11-20 |
CN111967494B true CN111967494B (en) | 2024-03-26 |
Family
ID=73361271
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010625497.1A Active CN111967494B (en) | 2020-07-01 | 2020-07-01 | Multi-source heterogeneous data analysis method for guard security of large movable public security system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111967494B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115409433B (en) * | 2022-11-02 | 2023-04-07 | 成都宏恒信息科技有限公司 | Depth NLP-based method and device for analyzing important community personnel portrait |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108536801A (en) * | 2018-04-03 | 2018-09-14 | 中国民航大学 | A kind of civil aviaton's microblogging security public sentiment sentiment analysis method based on deep learning |
CN108763542A (en) * | 2018-05-31 | 2018-11-06 | 中国华戎科技集团有限公司 | A kind of Text Intelligence sorting technique, device and computer equipment based on combination learning |
CN110472052A (en) * | 2019-07-31 | 2019-11-19 | 西安理工大学 | A kind of Chinese social platform sentiment analysis method based on deep learning |
-
2020
- 2020-07-01 CN CN202010625497.1A patent/CN111967494B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108536801A (en) * | 2018-04-03 | 2018-09-14 | 中国民航大学 | A kind of civil aviaton's microblogging security public sentiment sentiment analysis method based on deep learning |
CN108763542A (en) * | 2018-05-31 | 2018-11-06 | 中国华戎科技集团有限公司 | A kind of Text Intelligence sorting technique, device and computer equipment based on combination learning |
CN110472052A (en) * | 2019-07-31 | 2019-11-19 | 西安理工大学 | A kind of Chinese social platform sentiment analysis method based on deep learning |
Also Published As
Publication number | Publication date |
---|---|
CN111967494A (en) | 2020-11-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Yu et al. | Deep learning for real-time social media text classification for situation awareness–using Hurricanes Sandy, Harvey, and Irma as case studies | |
Vo et al. | Crime rate detection using social media of different crime locations and Twitter part-of-speech tagger with Brown clustering | |
Rawat et al. | Techniques for predicting dark web events focused on the delivery of illicit products and ordered crime | |
Al Nabki et al. | File Name Classification Approach to Identify Child Sexual Abuse. | |
Almehmadi et al. | Language usage on Twitter predicts crime rates | |
Hina et al. | Email classification and forensics analysis using machine learning | |
Pandya et al. | Analysis and Prediction of Location based Criminal Behaviors Through Machine Learning | |
Degadwala et al. | Location-Based Modelling And Analysis Of Threats By Using Text Mining | |
Siddiqui et al. | Criminal activity detection in social network by text mining: comprehensive analysis | |
Gautam et al. | A review on cyberstalking detection using machine learning techniques: Current trends and future direction | |
Krishnamoorthy et al. | A novel and secured email classification and emotion detection using hybrid deep neural network | |
CN111967494B (en) | Multi-source heterogeneous data analysis method for guard security of large movable public security system | |
CN110716957A (en) | Intelligent mining and analyzing method for class case suspicious objects | |
Norouzi | Spatial, temporal, and semantic crime analysis using information extraction from online news | |
CN116723005A (en) | Method and system for tracking malicious code implicit information under polymorphic hiding | |
Iqbal et al. | Artificial intelligence and digital forensics | |
Al-Shalabi | Perceptions of crime behavior and relationships: rough set based approach | |
Kshatri et al. | Crime Detection Approach Using Big Data Analytics and Machine Learning | |
Shaker et al. | Review on Analyzing and Detecting Crimes | |
CN114860903A (en) | Event extraction, classification and fusion method oriented to network security field | |
Hassan et al. | Prediction of terrorist activities by using unsupervised learning techniques | |
Yağanoğlu et al. | Separation of Incoming E-Mails Through Artificial Intelligence Techniques | |
Prasad et al. | Analysis and prediction of crime against woman using machine learning techniques | |
Boumahdi et al. | Crisis Management Systems: Big Data and Machine Learning Approach. | |
Saldaña | An empirical study of the perception of criminality through analysis of newspapers online |
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 |