Search

CN-121998382-A - Intelligent business location method, equipment and medium based on big data analysis

CN121998382ACN 121998382 ACN121998382 ACN 121998382ACN-121998382-A

Abstract

The invention discloses an intelligent business location method, equipment and medium based on big data analysis, which relate to the technical field of big data analysis and comprise the steps of collecting business location coupling data sets, executing feature alignment to form standard space-time data streams, adopting APACHE FLINK stream processing engines to execute CEP mode recognition on the standard space-time data streams to form reference point address digital images, inputting the reference point address digital images and the business location coupling data sets into a deep learning location model, executing time sequence modeling and space coding by a feature coding layer, performing space clustering analysis and similarity calculation by a similarity measurement layer, and outputting an address selection similarity cloud image. According to the invention, through APACHE FLINK stream processing engines and the deep learning site selection model, the fusion precision between multi-source data is enhanced, and the dynamic prediction capability of business site selection is remarkably improved.

Inventors

  • LI YINGLI

Assignees

  • 浙江可计算开店科技有限公司

Dates

Publication Date
20260508
Application Date
20260409

Claims (10)

  1. 1. An intelligent business location method based on big data analysis is characterized by comprising the following steps of, Acquiring a commercial site selection coupling data set, and executing characteristic alignment to form a standard space-time data stream, adopting APACHE FLINK stream processing engine to execute CEP mode identification on the standard space-time data stream to form a reference point address digital image; Inputting the digital representation of the reference point address and the commercial site selection coupling data set into a deep learning site selection model, performing time sequence modeling and spatial coding by a feature coding layer, performing spatial cluster analysis and similarity calculation by a similarity measurement layer, and outputting a site selection similarity cloud picture; performing region positioning according to the address selection similarity cloud image to form an alternative address set, performing multi-objective optimization on the alternative address set through a TOPSIS algorithm, performing collaborative filtering compensation by using a time-space attenuation factor, and generating a comprehensive score of the alternative address; And based on the comprehensive scores of the alternative addresses, sorting and single-hot encoding are carried out on the alternative address sets, a priority address selection list is obtained, cross verification is carried out on the priority address selection list, and an address selection decision package is output.
  2. 2. The intelligent business addressing method based on big data analysis of claim 1, wherein the business addressing coupling data set comprises three-dimensional addressing map, historical store operation data, user portrait type and business district activity data.
  3. 3. The intelligent business addressing method based on big data analysis of claim 2, wherein the forming of the digital representation of the reference point address comprises the steps of, Performing time stamp unification and space registration on the three-dimensional site selection map, the historical store operation data, the user portrait type and the business district active data to form a standard space-time data stream; And constructing a real-time data processing pipeline based on APACHE FLINK stream processing engines, inputting the standard space-time data stream into the real-time data processing pipeline for CEP mode identification, and outputting the digital representation of the reference point address.
  4. 4. The intelligent business addressing method based on big data analysis of claim 1, wherein the deep learning addressing model is specifically constructed as follows, Building a feature coding layer through a space-time convolution network, and building a similarity measurement layer by adopting a graph attention network; and carrying out gradient back propagation and cross-layer stacking on the feature coding layer and the similarity measurement layer, and constructing a deep learning site selection model.
  5. 5. The intelligent business addressing method based on big data analysis of claim 4, wherein the output addressing similarity cloud picture comprises the following steps, Inputting the digital representation of the reference point address and the commercial address selection coupling data set into a deep learning address selection model, and performing time sequence modeling and space coding by a feature coding layer through a space-time convolutional neural network to generate a dynamic index time sequence matrix; The similarity measurement layer performs spatial clustering analysis through a spectral clustering algorithm, and performs similarity calculation by adopting a cosine similarity formula to form a position similarity value; and performing spatial interpolation and thermodynamic diagram rendering on the dynamic index time sequence matrix and the position similarity value, and outputting an address selection similarity cloud picture.
  6. 6. The intelligent business addressing method based on big data analysis of claim 1, wherein the generating the composite score of the candidate address comprises the following steps, According to the site selection similarity cloud picture, carrying out hot spot detection and region positioning by using a region growing algorithm to form an alternative address set; performing attenuation correction and collaborative filtering compensation on the initial optimization score by using a time-space attenuation factor to generate a comprehensive score of the alternative address.
  7. 7. The intelligent business addressing method based on big data analysis of claim 6, wherein the acquiring the priority addressing list comprises the following steps, Sorting the candidate address sets based on the composite score of the candidate addresses to generate priority address classifications; and performing single-hot coding on the priority address classification to obtain a priority address selection list.
  8. 8. The intelligent business addressing method based on big data analysis of claim 1, wherein the output addressing decision package is specifically that a Bootstrap self-service sampling method is adopted to perform confidence cross-validation on a priority addressing list, and the addressing decision package is output.
  9. 9. A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the computer program is characterized in that the processor realizes the steps of the intelligent business addressing method based on big data analysis according to any one of claims 1-8 when executing the computer program.
  10. 10. A computer readable storage medium having a computer program stored thereon, wherein the computer program when executed by a processor performs the steps of the intelligent business addressing method based on big data analysis of any one of claims 1 to 8.

Description

Intelligent business location method, equipment and medium based on big data analysis Technical Field The invention relates to the technical field of big data analysis, in particular to an intelligent business site selection method, equipment and medium based on big data analysis. Background With the accelerated development of urbanization, commercial site selection is used as a core link of industries such as retail, catering, service and the like, and the scientificity and the precision of the commercial site selection directly influence the final operation efficiency and the market competitiveness. In recent years, big data analysis is widely applied to the business site selection decision support level as a technical means capable of integrating multi-source heterogeneous data and association relation mining. In addition, the existing GIS geographic information analysis method and machine learning model can also realize business district identification and potential demand prediction to a certain extent, and improve the intelligent level of business site selection. However, the prior art still has several bottlenecks in the practical application of commercial site selection. Firstly, insufficient space-time heterogeneity processing of multi-source data, lack of an effective feature alignment mechanism, and therefore difficulty in achieving accurate coupling between multi-source data, and insufficient representativeness of a finally determined site selection reference system. Secondly, in the process of similarity matching, the existing machine learning model is limited to static feature comparison, deep association between spatial clustering and time sequence evolution cannot be fully mined, and dynamic prediction capability of business site selection is limited. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. Therefore, the invention provides an intelligent commercial site selection method based on big data analysis, which solves the problems of insufficient representativeness of a site selection reference system and low commercial site selection prediction capability. In order to solve the technical problems, the invention provides the following technical scheme: In a first aspect, the invention provides an intelligent business addressing method based on big data analysis, which comprises the steps of collecting a business addressing coupling data set, executing characteristic alignment to form a standard space-time data stream, adopting a APACHE FLINK stream processing engine to execute CEP mode identification on the standard space-time data stream to form a reference point address digital image; Inputting the digital representation of the reference point address and the commercial site selection coupling data set into a deep learning site selection model, performing time sequence modeling and spatial coding by a feature coding layer, performing spatial cluster analysis and similarity calculation by a similarity measurement layer, and outputting a site selection similarity cloud picture; performing region positioning according to the address selection similarity cloud image to form an alternative address set, performing multi-objective optimization on the alternative address set through a TOPSIS algorithm, performing collaborative filtering compensation by using a time-space attenuation factor, and generating a comprehensive score of the alternative address; And based on the comprehensive scores of the alternative addresses, sorting and single-hot encoding are carried out on the alternative address sets, a priority address selection list is obtained, cross verification is carried out on the priority address selection list, and an address selection decision package is output. As a preferable scheme of the intelligent business location method based on big data analysis, the business location coupling data set comprises a three-dimensional location map, historical store operation data, user portrait types and business district activity data. As a preferable scheme of the intelligent business addressing method based on big data analysis, the invention comprises the following steps of forming a digital image of a reference point address, Performing time stamp unification and space registration on the three-dimensional site selection map, the historical store operation data, the user portrait type and the business district active data to form a standard space-time data stream; And constructing a real-time data processing pipeline based on APACHE FLINK stream processing engines, inputting the standard space-time data stream into the real-time data processing pipeline for CEP mode identification, and outputting the digital representation of the reference point address. As a preferable scheme of the intelligent business addressing method based on big data analysis, the deep learning addressing model is specifically constructed as