Communication method based on artificial intelligence

Abstract

The invention discloses a communication method based on artificial intelligence. The method comprises the steps of data acquisition and preprocessing, feature engineering and network architecture optimization. Collecting and preprocessing data; feature engineering: the feature engineering comprises feature selection, feature extraction and feature conversion, and a deep learning model is adopted to mine key factors influencing network performance and security in original data; network architecture optimization: performing intelligent analysis and management on network resources by using an intelligent algorithm, identifying bottlenecks and busy periods in the network by analyzing a large amount of real-time data, and predicting, simulating and adjusting resource allocation through the intelligent algorithm. The invention belongs to the technical field of communication. The technical effects of the invention are that the method combines artificial intelligence, predicts the network peak values in different time periods in real time, enables the communication network to have different peak flow values in different time periods, achieves the more precise and intelligent management of the communication network, and improves the performance and adaptability of the network.

Description

Technical Field

[0001] The present invention relates to the field of communication-related technologies, and in particular to a communication method based on artificial intelligence. Background Art

[0002] In the digital age, communication networks are the foundation of the information society, and their performance and efficiency are crucial to social life and economic development. With the continuous evolution of communication technology, network planning and optimization are particularly urgent to meet the growing needs of users and adapt to the complex communication environment.

[0003] Faced with the surge in data volume and the diversification of network service demands, traditional communication network management and optimization methods can no longer meet the needs of efficient and flexible operation and maintenance, especially when the number of users at different times varies greatly. If the maximum peak usage is set for network demand, network resources will be greatly wasted when there are only a small number of users. Even if the valley peak traffic is set according to daily usage needs, the communication network will easily become stuck during some special time periods. Summary of the invention

[0004] To this end, the present invention provides a communication method based on artificial intelligence to solve the above-mentioned problems in the prior art.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] According to a first aspect of the present invention, an artificial intelligence-based communication method includes data collection and preprocessing, feature engineering, and network architecture optimization;

[0007] Data collection and preprocessing: Build a multi-layer perceptron model, learn and classify the features of communication data, collect raw data, preprocess the raw data, clean it, remove duplicates, remove erroneous data, and delete irrelevant fields, and convert the classified data into numerical data and normalize the numerical data;

[0008] Feature engineering: Feature engineering includes feature selection, feature extraction, and feature conversion. It uses deep learning models to mine the key factors that affect network performance and security in the original data.

[0009] Network architecture optimization: Use intelligent algorithms to intelligently analyze and manage network resources. By analyzing large amounts of real-time data, bottlenecks and busy periods in the network can be identified. Through prediction and simulation by intelligent algorithms, resource allocation can be adjusted.

[0010] Furthermore, the goals of raw data collection include network equipment logs, user behavior data, network performance indicators, network security data, network fault data, network configuration data, and network service quality.

[0011] Furthermore, deep learning models used in feature engineering include: Convolutional neural networks: extract spatial features for processing images and spatial data;

[0012] Recurrent Neural Network: Processes sequential data and is used to solve time series and traffic forecasting problems.

[0013] Furthermore, the network architecture optimization includes the following steps:

[0014] S1: Demand analysis and planning: evaluate the types of services, data volume, user density, and service quality requirements that the network needs to support, and plan the network's coverage, capacity, equipment type, and configuration based on the demand analysis results;

[0015] S2: Time planning and deployment, through the prediction and real-time decision-making of convolutional neural networks and recurrent neural networks, the network realizes dynamic spectrum allocation in different time periods and makes intelligent adjustments based on expected user needs and network status;

[0016] S3: Network simulation: building a network simulation model to simulate network performance under different time deployment schemes;

[0017] S4: Monitor network performance in real time, use the Simple Network Management Protocol to collect performance data of network devices and services, deploy sensors on key nodes of the network, and collect network performance indicators in real time;

[0018] S5: Threat detection, using graph neural networks to achieve real-time detection of known and unknown threats, capturing the complex interactions and dynamic evolution between network entities, and proactively identifying and blocking potential attack behaviors before threats occur;

[0019] S6: Network optimization, first use convolutional neural networks and recurrent neural networks to continuously monitor the status and communication quality of network devices, identify abnormal behavior in real time, and use self-organizing networks to automatically adjust system settings based on previous successful repair cases added when training the model.

[0020] Furthermore, the time periods in S2 include 8:00-12:00, 12:00-14:00, 14:00-18:00, 18:00-23:00, and 23:00-8:00 of the next day.

[0021] Furthermore, the network performance indicators in S4 include flow rate, delay rate, and packet loss rate.

[0022] Furthermore, the packet loss rate calculation formula is:

[0023]

[0024] Where: PLR is the packet loss rate, L is the number of packets lost in the network connection, and N is the total number of packets sent in the network connection.

[0025] Furthermore, the delay rate calculation formula is:

[0026]

[0027] Where: Y i is the delay time in each network transmission, V i is the average value of all measured delays, N is the number of delay measurements, and L is the total number of delays.

[0028] The present invention has the following advantages:

[0029] The present invention combines artificial intelligence to predict network peaks in different time periods in real time, so that the communication network has different peak flows in different time periods, thereby achieving more accurate and intelligent management of the communication network and improving the performance and adaptability of the network. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A flow chart of an artificial intelligence-based communication method provided by the present invention.

[0031] Figure 2 A flowchart of network architecture optimization of an artificial intelligence-based communication method provided by the present invention. DETAILED DESCRIPTION

[0032] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] See also Figure 1 and Figure 2 As shown, the present invention provides an artificial intelligence-based communication method, including data collection and preprocessing, feature engineering, and network architecture optimization;

[0034] Data collection and preprocessing

[0035] By building a multi-layer perceptron (MLP) model, the communication data is feature learned and classified, and the original data is collected;

[0036] The goals of raw data collection include network equipment logs, user behavior data, network performance indicators, network security data, network fault data, network configuration data, and network service quality;

[0037] The raw data is preprocessed, cleaned, deduplicated, erroneous data removed, and irrelevant fields deleted to ensure the quality and accuracy of the data, and the categorical data is converted into numerical data and the numerical data is normalized so that the artificial intelligence algorithm can better process the data. Specifically:

[0038] Data cleaning: remove duplicates and invalid items in the data, such as null values, invalid characters, etc., and correct errors in the data, such as format errors, logical errors, etc.;

[0039] Data deduplication: Use deduplication algorithms or tools to deduplicate the collected data to ensure the uniqueness of the data;

[0040] Removal of irrelevant fields: According to the analysis requirements, delete the fields irrelevant to the analysis target to reduce the redundancy and complexity of the data;

[0041] Data normalization: converting categorical data into numerical data; normalizing or standardizing numerical data so that the data can be compared and analyzed on a uniform scale;

[0042] Data integration: Integrate the data after cleaning, deduplication, deletion of irrelevant fields and normalization to form a complete data set;

[0043] Data verification: Use data verification tools or algorithms to verify and validate the processed data to ensure data integrity and consistency;

[0044] The MLP model can also be trained using labeled abnormal traffic samples and normal traffic samples, and the model parameters can be continuously optimized through the back-propagation algorithm, thus facilitating the subsequent accurate collection of raw data;

[0045] Feature Engineering

[0046] Feature engineering includes feature selection, feature extraction, and feature conversion to mine the key factors that affect network performance and security in the original data;

[0047] CNN can be used to extract spatial features, which is suitable for processing images and spatial data, such as network geographic location and IP address. RNN can be used to process sequential data, which is suitable for solving problems such as time series and traffic prediction, such as user behavior data.

[0048] Specifically:

[0049] Feature Selection

[0050] Correlation analysis: Use statistical methods (correlation coefficient, mutual information) to evaluate the correlation between each feature and the target variable.

[0051] Feature Importance Assessment: Use the feature importance assessment tool to determine which features have a significant impact on model performance.

[0052] Feature selection strategy: Based on the results of correlation analysis and feature importance assessment, select the features that contribute most to model performance;

[0053] Feature extraction

[0054] Use machine learning algorithms (CNN, RNN) to automatically extract features, such as statistical features of network traffic (such as packet size, transmission rate, etc.) and pattern features of user behavior (such as login frequency, access pattern, etc.);

[0055] Feature Transformation

[0056] Scaling and Normalization: Features are scaled and normalized to ensure that all features are at the same scale.

[0057] Encoding: Encoding categorical features helps convert them into numerical forms that the model can process.

[0058] Polynomial features and interaction features: Generate polynomial features and interaction features to capture complex relationships between features;

[0059] Network architecture optimization

[0060] S1: Demand analysis and planning. Evaluate the types of services, data volume, user density, and quality of service (QoS) requirements that the network needs to support. Based on the demand analysis results, plan the network's coverage, capacity, equipment type, and configuration;

[0061] S2: Time planning and deployment. Through the prediction and real-time decision-making of convolutional neural networks (CNN) and recurrent neural networks (RNN), the network realizes dynamic spectrum allocation in different time periods and makes intelligent adjustments based on expected user needs and network status. The time periods include 8:00-12:00, 12:00-14:00, 14:00-18:00, 18:00-23:00, and 23:00-8:00 the next day.

[0062] S3: Network simulation, building a network simulation model to simulate network performance under different time deployment scenarios. During the simulation process, key performance indicators such as signal strength, data transmission rate, and latency are tracked;

[0063] S4: Monitor network performance in real time, use the Simple Network Management Protocol (SNMP) to collect performance data of network devices and services, deploy sensors at key nodes of the network, and collect information such as traffic, latency, and packet loss in real time;

[0064] The frequency of a specific data packet being lost during transmission is calculated as:

[0065]

[0066] Where: PLR is the packet loss rate, which is the ratio of the number of lost packets to the total number of sent packets, L is the number of packets lost in the network connection, and N is the total number of packets sent in the network connection.

[0067] The rate of change of data delay reflects the stability of the network, which is calculated as:

[0068]

[0069] Where: Y i is the delay time in each network transmission, V i is the average value of all measured delays, N is the number of delay measurements, and L is the total number of delays;

[0070] S5: Threat detection, using deep learning, ensemble learning and other algorithms; achieving real-time detection of known and unknown threats. Graph Neural Network (GNN) can capture the complex interactions and dynamic evolution between network entities, proactively identify and block potential attack behaviors before the threat occurs, thereby minimizing the impact scope and loss of security incidents, and achieving early warning of potential threats;

[0071] S6: Network optimization, first use CNN and RNN to continuously monitor the status and communication quality of network devices, and identify abnormal behaviors in real time, such as a sudden drop in signal strength or interruption in data transmission. Once these anomalies are detected, the system can quickly locate the source of the problem and use self-organizing networks (SON) to automatically adjust system settings based on previous successful repair cases added when training the model to adapt to different systems and scenarios, thereby improving the efficiency and performance of the network and automatically recommending or implementing repair measures.

Claims

1. A communication method based on artificial intelligence, characterized in that: Including data collection and preprocessing, feature engineering, and network architecture optimization; Data collection and preprocessing: Build a multi-layer perceptron model, learn and classify the features of communication data, collect raw data, preprocess the raw data, clean it, remove duplicates, remove erroneous data, and delete irrelevant fields, and convert the classified data into numerical data and normalize the numerical data; Feature engineering: Feature engineering includes feature selection, feature extraction, and feature conversion. It uses deep learning models to mine the key factors that affect network performance and security in the original data. Network architecture optimization: Use intelligent algorithms to intelligently analyze and manage network resources. By analyzing large amounts of real-time data, bottlenecks and busy periods in the network can be identified. Through prediction and simulation by intelligent algorithms, resource allocation can be adjusted.

2. The communication method based on artificial intelligence according to claim 1, characterized in that: The goals of raw data collection include network equipment logs, user behavior data, network performance indicators, network security data, network fault data, network configuration data, and network service quality.

3. The communication method based on artificial intelligence according to claim 1, characterized in that: The deep learning models used in feature engineering include: Convolutional neural networks: extract spatial features for processing images and spatial data; Recurrent Neural Network: Processes sequential data and is used to solve time series and traffic forecasting problems.

4. The communication method based on artificial intelligence according to claim 1, characterized in that: Network architecture optimization includes the following steps: S1: Demand analysis and planning: evaluate the types of services, data volume, user density, and service quality requirements that the network needs to support, and plan the network's coverage, capacity, equipment type, and configuration based on the demand analysis results; S2: Time planning and deployment, through the prediction and real-time decision-making of convolutional neural networks and recurrent neural networks, the network realizes dynamic spectrum allocation in different time periods and makes intelligent adjustments based on expected user needs and network status; S3: Network simulation: building a network simulation model to simulate network performance under different time deployment schemes; S4: Monitor network performance in real time, use the Simple Network Management Protocol to collect performance data of network devices and services, deploy sensors on key nodes of the network, and collect network performance indicators in real time; S5: Threat detection, using graph neural networks to achieve real-time detection of known and unknown threats, capturing the complex interactions and dynamic evolution between network entities, and proactively identifying and blocking potential attack behaviors before threats occur; S6: Network optimization, first use convolutional neural networks and recurrent neural networks to continuously monitor the status and communication quality of network devices, identify abnormal behavior in real time, and use self-organizing networks to automatically adjust system settings based on previous successful repair cases added when training the model.

5. The communication method based on artificial intelligence according to claim 4, characterized in that: The time periods in S2 include 8:00-12:00, 12:00-14:00, 14:00-18:00, 18:00-23:00, and 23:00-8:00 of the next day.

6. The communication method based on artificial intelligence according to claim 4, characterized in that: The network performance indicators in S4 include traffic, delay rate, and packet loss rate.

7. The communication method based on artificial intelligence according to claim 6, characterized in that: The packet loss rate calculation formula is: Where: PLR is the packet loss rate, L is the number of packets lost in the network connection, and N is the total number of packets sent in the network connection.

8. The communication method based on artificial intelligence according to claim 6, characterized in that: The calculation formula of delay rate D is: Where: Y i is the delay time in each network transmission, V i is the average value of all measured delays, N is the number of delay measurements, and L is the total number of delays.

Images