Lightweight time sequence compression method with adaptive segmentation optimization
Through the lightweight time series compression method of adaptive segmentation optimization, dynamic segmentation processing and heuristic search algorithm optimization parameters, the problem of large delay in the time series data compression algorithm is solved, and efficient data transmission and storage is achieved.
Patent Information
- Application Number
- CN202510603996.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-12
- Publication Date
- 2025-08-26
AI Technical Summary
Existing time series data compression algorithms have high latency when processing complex and changing data, and the transmission pressure is too high, making it difficult to meet the needs of low latency and efficient data transmission.
Adaptive segmentation optimization lightweight time series compression method is adopted to generate fluctuating feature sequences through dynamic segmentation processing, combined with the heuristic search algorithm to optimize parameters, dynamic selection compression algorithm is used to encode, and efficient encoding is achieved using leading zero and trailing zero characteristics.
It significantly improves the compression ratio and decompression speed, reduces hardware resource consumption, and can handle real-time compression tasks of millions of data points, and is suitable for IoT sensor networks and real-time financial analysis scenarios.
Smart Images
Figure CN120546699A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of lossless compression of time series data, and in particular relates to a lightweight time series compression method with adaptive segmentation optimization. Background Art
[0002] In this era of data explosion, data compression has attracted widespread attention. The rise of IoT devices is driving a massive surge in data. This data includes numerous continuous time-series outputs, such as sensor data (temperature, humidity, etc.) within the IoT. How should the present invention handle such a large volume of data transmission? Without processing this data, the present invention's network would struggle to withstand the high traffic pressure, leading to problems such as excessive transmission delays. Therefore, there is an urgent need to find a method to reduce data transmission pressure while maintaining data speed. As we all know, data compression is a very common method, originally used for compressing files and other applications. If the present invention applies data compression appropriately to time series data, it will effectively improve data transmission and storage efficiency. Compression is a key step. Today's compression algorithms are generally able to effectively utilize features in the data, such as timestamps in time series data. Compression algorithms can also effectively compress time series data because they can design different algorithmic strategies for different data patterns and regularities. The choice of compression algorithm is crucial, and low latency is crucial for time series processing systems. Furthermore, the effectiveness of compression algorithms has been demonstrated through various case studies. However, the encoding time associated with compression techniques often results in significant latency. For example, algorithms such as Gzip and LZ4 have been shown to have significant time overhead in experiments. Therefore, the present invention requires an algorithm that can adaptively adjust the input time series data while achieving excellent compression ratios and maintaining acceptable time overhead. In summary, effectively processing compressed time series data presents three key challenges. First, the present invention requires a suitable compression algorithm to handle complex and constantly changing data. Second, time series data is continuously transmitted or stored, requiring the compression algorithm to dynamically adapt to changes in data characteristics. Third, the huge amount of time series data requires the use of appropriate processing methods to adapt to the compression algorithm. Summary of the Invention
[0003] To solve the above technical problems, the present invention proposes a lightweight time series compression method with adaptive segmentation optimization to solve the problems existing in the above-mentioned prior art.
[0004] In a first aspect, to achieve the above-mentioned object, the present invention provides a lightweight time series compression method with adaptive segmentation optimization, comprising the following steps:
[0005] S1. Dynamically segment the input time series floating-point data stream. By calculating the binary XOR values of adjacent data points to generate a fluctuation feature sequence, the data fluctuation degree is determined based on an adaptive threshold, and the data stream is divided into several segmented units with stable features.
[0006] S2. Parameter optimization is performed on each segment unit. The initial optimization space is constructed using the fluctuation characteristics of the segment data. The optimal leading zero coding matrix is iteratively selected through a heuristic search algorithm to adapt the compression algorithm to the data distribution characteristics of the current segment.
[0007] S3. Based on the fluctuation characteristics and repeatability characteristics of the optimized segments, an adaptive compression algorithm is dynamically selected for encoding processing. High-fluctuation segments utilize the leading and trailing zero characteristics to achieve efficient encoding, and high-repetition segments utilize indexing and storage mechanisms to optimize the compression of repeated patterns.
[0008] Optionally, the dynamic segmentation processing in S1 includes:
[0009] Convert the original floating-point data into a binary sequence and calculate the XOR value of adjacent data bits The mean μ and standard deviation σ of the L_count change rate are calculated through a sliding window. When the L_count change of two consecutive data points exceeds μ+σ, it is determined as a fluctuation mutation point, which is used to divide the segment boundary.
[0010] Optionally, the parameter optimization process in S2 includes:
[0011] A 7-bit binary coding space with L_count value as gene is established, and an elite retention strategy genetic algorithm is used for iterative optimization. The fitness function is set to C=C1+C2. The convergence is accelerated by dynamically adjusting the crossover probability and mutation probability, and finally the optimal leading zero matrix containing 7 zero bits is output.
[0012] Optionally, the algorithm selection strategy in S3 includes:
[0013] The L_count standard deviation σ_L and the repeatability index ρ_repeat of the segmented data are calculated. When σ_L> the preset threshold τ_v, the ACTF variable-length coding module is activated. When ρ_repeat> the preset threshold τ_r, the ACTF128 fixed-length coding module is activated. The threshold setting method is: τ_v=std(diff(len(L_count change))), τ_r=mean(diff(len(L_count change))). The dual-mode selection mechanism improves the compression ratio efficiency by 37%.
[0014] Optionally, during the dynamic segmentation process, the calculation of the adaptive threshold includes weighted correction of the data trend, specifically threshold = μ_L + σ_L*T a (T a is the threshold adjustment range you set).
[0015] Optionally, constructing the initial population in the heuristic search algorithm includes: removing repeated L_count values and taking the first 7 unique values as parents, and generating the remaining individuals through gene recombination. The recombination operation adopts a combination of single-point crossover and uniform mutation. The crossover probability is initialized to 0.8 and the mutation probability is initialized to 0.1.
[0016] Optionally, in the compression algorithm selection module, the ACTF algorithm utilizes the characteristics of leading zeros and trailing zeros to achieve efficient encoding. This encoding strategy improves the compression efficiency of high-volatility data by 53%.
[0017] Optionally, in the compression algorithm selection module, the ACTF128 algorithm has a built-in index dictionary containing 128 historical index values, and the dictionary update mechanism adopts a sliding window frequency statistics method.
[0018] In a second aspect, the present invention further provides a computer terminal device, comprising:
[0019] one or more processors;
[0020] a memory, coupled to the processor, for storing one or more programs;
[0021] When the one or more programs are executed by the one or more processors, the one or more processors implement a lightweight time series compression method with adaptive segmentation optimization.
[0022] In a third aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a lightweight time series compression method with adaptive segmentation optimization.
[0023] Compared with the prior art, the present invention has the following advantages and technical effects:
[0024] The present invention provides a lightweight time series compression method with adaptive segmentation optimization, which achieves a comprehensive improvement in compression performance through dynamic segmentation, parameter optimization and dual-algorithm collaborative mechanism. First, at the data segmentation level, a dynamic segmentation strategy based on L_count fluctuation characteristics is adopted. The data mutation point is determined by adaptive threshold, and the time series is divided into segmented units with stable characteristics. Compared with the traditional fixed threshold method, it effectively adapts to the dynamic changes of data flow patterns. Secondly, in the parameter optimization link, the heuristic genetic algorithm is innovatively combined with the 7-bit binary coding space. By dynamically adjusting the crossover / mutation probability to accelerate convergence, the leading zero matrix optimization efficiency is improved by 37%, the hardware resource consumption is reduced by 68%, and the compression algorithm's adaptability to local data features is significantly improved. On this basis, the framework further constructs a dual-algorithm collaborative mechanism. By real-time analysis of the volatility and repetitive characteristics of the segmented data, it dynamically switches between ACTF variable-length entropy coding and ACTF128 fixed-length pattern matching strategies: for highly volatile data, a 53% compression efficiency improvement is achieved; for highly repetitive data, the sliding window frequency statistics method is used to dynamically update the ACTF128 index dictionary. Industrial-grade testing has demonstrated that WZS supports 10Gbps high-speed data stream processing, with end-to-end latency kept below 5ms and decompression speeds 16.45 times faster than traditional methods. This technology has been validated in IoT sensor networks and real-time financial analysis scenarios, reliably handling real-time compression tasks involving millions of data points, providing an efficient solution for low-latency, high-density data transmission. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The accompanying drawings, which constitute part of the present invention, are provided to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are provided to explain the present invention and do not constitute an undue limitation of the present invention. In the accompanying drawings:
[0026] Figure 1 This is a schematic diagram of a lightweight time series compression framework structure in a lightweight time series compression method with adaptive segmentation optimization in an embodiment of the present invention. DETAILED DESCRIPTION
[0027] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0028] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0029] like Figure 1As shown, this embodiment provides a lightweight time series compression method with adaptive segmentation optimization, including:
[0030] S1. Dynamically segment the input time series floating-point data stream. By calculating the binary XOR values of adjacent data points to generate a fluctuation feature sequence, the data fluctuation degree is determined based on an adaptive threshold, and the data stream is divided into several segmented units with stable features.
[0031] S2. Parameter optimization is performed on each segment unit. The initial optimization space is constructed using the fluctuation characteristics of the segment data. The optimal leading zero coding matrix is iteratively selected through a heuristic search algorithm to adapt the compression algorithm to the data distribution characteristics of the current segment.
[0032] S3. Based on the fluctuation characteristics and repeatability characteristics of the optimized segments, an adaptive compression algorithm is dynamically selected for encoding processing. High-fluctuation segments utilize the leading and trailing zero characteristics to achieve efficient encoding, and high-repetition segments utilize indexing and storage mechanisms to optimize the compression of repeated patterns.
[0033] Furthermore, the dynamic segmentation process in S1 includes:
[0034] Convert the original floating-point data into a binary sequence and calculate the XOR value of adjacent data bits The mean μ and standard deviation σ of the L_count change rate are calculated through a sliding window. When the L_count change of two consecutive data points exceeds μ+σ, it is determined as a fluctuation mutation point, which is used to divide the segment boundary.
[0035] Furthermore, the parameter optimization process in S2 includes:
[0036] A 7-bit binary coding space with L_count value as gene is established, and an elite retention strategy genetic algorithm is used for iterative optimization. The fitness function is set to C=C1+C2. The convergence is accelerated by dynamically adjusting the crossover probability and mutation probability, and finally the optimal leading zero matrix containing 7 zero bits is output.
[0037] Furthermore, the algorithm selection strategy in S3 includes:
[0038] The L_count standard deviation σ_L and the repeatability index ρ_repeat of the segmented data are calculated. When σ_L> the preset threshold τ_v, the ACTF variable-length coding module is activated. When ρ_repeat> the preset threshold τ_r, the ACTF128 fixed-length coding module is activated. The threshold setting method is: τ_v=median(σ_L)+IQR(σ_L)1.5, τ_r=mean(ρ_repeat)+2std(ρ_repeat). The dual-mode selection mechanism improves the compression ratio efficiency by 37%.
[0039] Furthermore, in the dynamic segmentation process, the calculation of the adaptive threshold includes weighted correction of the data trend, specifically threshold = μ_L + σ_L*T a (T a is the threshold adjustment range you set).
[0040] Furthermore, the heuristic search algorithm constructs the initial population including: removing duplicate L_count values and taking the first 7 unique values as parents, and generating the remaining individuals through gene recombination. The recombination operation adopts a combination of single-point crossover and uniform mutation. The crossover probability is initialized to 0.8 and the mutation probability is initialized to 0.1.
[0041] Furthermore, in the compression algorithm selection module, the ACTF algorithm uses the characteristics of leading zeros and trailing zeros to achieve efficient encoding. This encoding strategy improves the compression efficiency of high-volatility data by 53%.
[0042] Furthermore, in the compression algorithm selection module, the ACTF128 algorithm has a built-in index dictionary, which contains 128 historical index values, and the dictionary update mechanism adopts a sliding window frequency statistics method.
[0043] Specifically, this invention introduces WZS, a compression-based processing framework designed to overcome these three challenges. First, to meet the demand for lightweight and fast compression algorithms in time series data processing, ensuring low latency and excellent compression ratios, WZS uses a lightweight compression method to improve efficiency for various time series data types. This is a highly effective compression algorithm developed by this invention. Second, WZS includes a parameter optimizer, which will be described later in this invention. Third, WZS includes a module for processing time series data in segments. When the data format fluctuates or changes, it automatically detects and records the location of the change. It adaptively calculates the optimal parameters within the data segment, providing optimal performance advantages for time series data with different characteristics. The stream processing module in this framework divides the data into blocks based on the change in data characteristics between the current and previous change points. Each block is characterized by features extracted by the parameter optimizer, and the final iteration of the parameter optimizer produces an optimal leading zero matrix compatible with the compression algorithm of this invention. Once the compression algorithm obtains this leading zero matrix, it compresses all data within the block to achieve the optimal compression ratio for that segment. This method aims to simulate the dynamics of data streaming using time series data. By capturing significant fluctuations in real-time data, the present invention can immediately extract valuable information and optimize the compression performance of the corresponding blocks.
[0044] The present invention involves three modules, namely a data segmentation module, a segmentation optimization module, and an algorithm selection module.
[0045] First is the data segmentation module. The data segmentation module of the present invention analyzes the data set through a series of calculations and statistical evaluations. Each floating point value is marked as D = {d1, d2, ..., d n}, converted to binary (expressed as b n =binary(d n )), which is then expressed as B = {b1, b2, ... b n The present invention retains the initial data points, and the converted binary data undergoes XOR operation before and after processing (expressed as ), we get (x1=b1). Then, this value is used to calculate L_count. The formula for calculating L_count is expressed as: The present invention evaluates the L_count values of previous and subsequent data points. If the difference between these values exceeds an adaptive threshold dynamically calculated by the present invention, the present invention classifies the data as exhibiting significant fluctuation. This adaptive threshold is derived from the mean and standard deviation of the L_count variation. When such a fluctuation is detected, the present invention designates the current point as the starting point and the next key fluctuation as the end point, splitting the data between these points for further optimization. The present invention recognizes that adaptive thresholding is essential to the algorithm even in simple implementations because it eliminates the constraints of a fixed threshold and ensures robust performance across different data sets. Furthermore, L_count is a key component of the present invention's compression strategy, providing valuable insights for subsequent processing. It not only aids in parameter optimization but also plays a key role in selecting the most appropriate compression algorithm.
[0046] In data optimization, after obtaining the L_count information from the data segmentation module, the data enters the parameter optimization stage. At this stage, the present invention specifically optimizes the data segment for the subsequent compression algorithm, because L_count is a key parameter to guide the execution of the compression strategy. Therefore, a parameter optimizer is used to iteratively refine the optimal leading zero matrix within the data segment and integrate it into the compression algorithm to ultimately achieve the optimal compression ratio of the data segment. In the parameter optimizer, the concept of heuristic genetic algorithm must be combined. First, the parameter optimizer uses the L_count information in the segmented data to construct an initial population. Here, the present invention assumes that the set L_count in the segmentation function is S = {L1, L2, ..., L n}, where Ln represents the L of n data segments. After removing duplicates, a unique set U = {u1,u2,...,u m}, m<=n, randomly select 7 elements from it, expressed as V={v1,v2,...,v7},|V|=7. From this we can infer that i n =[g1,g2,...,g 63 ],g n∈{0,1}. Then, data processing is performed, such as removing duplicate L_count and selecting 7 unique L_count. This step aims to directly assign L_count as coordinates to in, with each L_count corresponding to a zero, and the value of in is as follows: At the same time, the present invention imposes certain constraints on the initial population and offspring during the iteration process to achieve the ultimate goal of the present invention: (make sure there are only 7 zeros), (Range from 0 to 62), the reason for using 7 zeros is that the subsequent compression algorithm allocates 3 bits for each L_count. In order to effectively use these 3 bits, L_count must be expressed as a binary value, such as "001", "011", etc., to reflect the corresponding number of L_count. For example, if V1=2, then i[2]=0, and its binary code is "010". The present invention improves the fitness function by abandoning the previous penalty mechanism and using positive incentives to avoid discontinuous fitness values. When the current "0" position of the present invention is between the set minimum and maximum values, a positive incentive mechanism will be given, and the formula is as follows,
[0047]
[0048] Where I(@) is an indicator function, which is 1 if the condition is met and 0 otherwise.
[0049] Likewise, the distance between adjacent “0”s is also within the constraints imposed by the present invention, and the corresponding positive excitation is given as follows:
[0050]
[0051] Then, the present invention returns the fitness function value. At the same time, the fitness function value of the present invention is the reciprocal of the number that meets the conditions, so the smaller the fitness value, the better the result of the present invention, C=C1+C2, Finally, after completing the specified number of iterations, the present invention performs the selection, crossover, and mutation steps. During this process, the present invention introduces dynamic parameter adjustment, which mainly adjusts the crossover and mutation probabilities by calculating the diversity (standard deviation) of the population, according to the following formula (k is the adjustment coefficient).
[0052]
[0053] This method helps to more effectively guide the iteration to reach the optimal result. Ultimately, the present invention can obtain the best fitness value. At the end of the parameter optimization process, the optimal leading zero parameter matrix corresponding to the segment with the best fitness value will be obtained.
[0054] In the compression algorithm selection module, for the two algorithms in the framework, the present invention has developed an efficient and low-cost switch that can automatically select a more suitable algorithm (ACTF or ACTF128) according to data characteristics (such as volatility, repeatability, L_count, etc.). It can be seen in the present invention that the selector determines which algorithm to select by calculating the standard deviation of L_count in the data segment to obtain its volatility and repeatability in the data segment. The present invention sets corresponding volatility and repeatability thresholds based on the characteristics of the two algorithms. When the volatility is greater than the threshold, the present invention selects the ACTF algorithm, and when the repeatability is greater than the threshold, the present invention selects the ACT F128 algorithm. By using a simple selector, it can not only be used as a selection algorithm, but also greatly reduce the cost of computing resources.
[0055] Experiments were conducted on 33 different datasets. Compared to time series lossless compression algorithms such as Gorilla and Chimp, WZS achieved an average 36.41% improvement in compression ratio, with a peak improvement of 53.08%. Compared to general algorithms, WZS achieved an average 15.07% improvement in compression ratio. Furthermore, WZS reduced compression and decompression time by two orders of magnitude compared to general algorithms and was 16.45% faster than other floating-point lossless time series compression methods. The experimental results, shown in Table 1, demonstrate that WZS excels in processing large-scale and variable time series data, significantly improving upon existing algorithms.
[0056] Table 1
[0057]
[0058] We developed a novel compression framework specifically designed for various time series data scenarios. This framework integrates two algorithms and includes an algorithm selector, achieving excellent compression ratios and efficiency on test datasets.
[0059] The present invention designs a data segmentation module that can segment a time series floating point data stream into local segments, thereby optimizing compression efficiency. In addition, the module allows the time series compression algorithm to work similarly to the stream compression algorithm.
[0060] The present invention develops a parameter optimizer that can locally optimize segmented data while extracting and refining information from datasets with different characteristics. This enhancement enables the compression algorithm to dynamically adapt and achieve superior performance based on the extracted information.
[0061] The present invention also provides a computer terminal device, comprising:
[0062] one or more processors;
[0063] a memory, coupled to the processor, for storing one or more programs;
[0064] When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement steps of a method such as XX.
[0065] The present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of a lightweight time series compression method with adaptive segmentation optimization are implemented.
[0066] In this embodiment, an electronic device is further provided, including a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to execute the method in the above embodiment.
[0067] The above program can be run in the processor, or it can be stored in the memory (or computer-readable medium), which includes permanent and non-permanent, removable and non-removable media and can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.
[0068] These computer programs can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps of the functions specified in one or more blocks can be implemented by different modules corresponding to different steps.
[0069] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.
Claims
1. A lightweight time series compression method with adaptive segmentation optimization, characterized in that The following steps are involved: S1. Dynamically segment the input time series floating-point data stream. By calculating the binary XOR values of adjacent data points to generate a fluctuation feature sequence, the data fluctuation degree is determined based on an adaptive threshold, and the data stream is divided into several segmented units with stable features. S2. Parameter optimization is performed on each segment unit. The initial optimization space is constructed using the fluctuation characteristics of the segment data. The optimal leading zero coding matrix is iteratively selected through a heuristic search algorithm to adapt the compression algorithm to the data distribution characteristics of the current segment. S3. Based on the fluctuation characteristics and repeatability characteristics of the optimized segments, an adaptive compression algorithm is dynamically selected for encoding processing. High-fluctuation segments utilize the leading and trailing zero characteristics to achieve efficient encoding, and high-repetition segments utilize indexing and storage mechanisms to optimize the compression of repeated patterns.
2. The lightweight time series compression method with adaptive segmentation optimization according to claim 1, characterized in that: The dynamic segmentation process in S1 includes: Convert the original floating-point data into a binary sequence and calculate the XOR value of adjacent data bits The mean μ and standard deviation σ of the L_count change rate are calculated through a sliding window. When the L_count change of two consecutive data points exceeds μ+σ, it is determined as a fluctuation mutation point, which is used to divide the segment boundary.
3. The lightweight time series compression method with adaptive segmentation optimization according to claim 1, characterized in that: The parameter optimization process in S2 includes: A 7-bit binary coding space with L_count value as gene is established, and an elite retention strategy genetic algorithm is used for iterative optimization. The fitness function is set to C=C1+C2. The convergence is accelerated by dynamically adjusting the crossover probability and mutation probability, and finally the optimal leading zero matrix containing 7 zero bits is output.
4. The lightweight time series compression method with adaptive segmentation optimization according to claim 1, characterized in that: The algorithm selection strategy in S3 includes: The L_count standard deviation σ_L and the repeatability index ρ_repeat of the segmented data are calculated. When σ_L> the preset threshold τ_v, the ACTF variable-length coding module is activated. When ρ_repeat> the preset threshold τ_r, the ACTF128 fixed-length coding module is activated. The threshold setting method is: τ_v=std(diff(len(L_count change))), τ_r=mean(diff(len(L_count change))). The dual-mode selection mechanism improves the compression ratio efficiency by 37%.
5. The lightweight time series compression method with adaptive segmentation optimization according to claim 1, characterized in that: In the dynamic segmentation process, the calculation of the adaptive threshold includes weighted correction of the data trend, specifically threshold = μ_L + σ_L*T a (T a is the threshold adjustment range you set).
6. The lightweight time series compression method with adaptive segmentation optimization according to claim 1, characterized in that: The initial population construction in the heuristic search algorithm includes: removing duplicate L_count values and taking the first 7 unique values as parents. The remaining individuals are generated by gene recombination. The recombination operation adopts a combination of single-point crossover and uniform mutation. The crossover probability is initialized to 0.8 and the mutation probability is initialized to 0.
1.
7. The lightweight time series compression method with adaptive segmentation optimization according to claim 1, characterized in that: In the compression algorithm selection module, the ACTF algorithm uses the characteristics of leading zeros and trailing zeros to achieve efficient encoding. This encoding strategy improves the compression efficiency of high-volatility data by 53%.
8. The lightweight time series compression method with adaptive segmentation optimization according to claim 1, characterized in that: In the compression algorithm selection module, the ACTF128 algorithm has a built-in index dictionary, which contains 128 historical index values, and the dictionary update mechanism adopts a sliding window frequency statistics method.
9. A computer terminal device, characterized in that: include: one or more processors; a memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the lightweight time series compression method with adaptive segmentation optimization as described in any one of claims 1 to 8.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the lightweight time series compression method with adaptive segmentation optimization according to any one of claims 1 to 8 is implemented.