Residential replacement matching system based on information analysis

Abstract

The invention relates to a residence replacement matching system based on information analysis. The residence replacement matching system comprises a picture feature processing subsystem, a user demand analysis module, a replacement matching index module and a matching preference screening module. According to the system, concrete house characteristics are generated through house VR images, and precise replacement requirements are generated in combination with user information; house matching information is efficiently screened according to the index priorities of the demand items, and an evaluation value output result is calculated through an evaluation algorithm containing user preferences. According to the invention, the problems of difficult demand acquisition and low matching granularity in the prior art can be solved, the matching efficiency and accuracy are improved, the housing resource matching reliability is guaranteed, and the living replacement demand of the user can be better met.

Description

Technical Field

[0001] This invention relates to residential information processing technology, and more specifically, to a residential replacement matching system based on information analysis. Background Technology

[0002] With the personalization and diversification of housing needs, housing replacement matching systems based on information analysis have gradually become an important research direction in the field of housing management and optimization. However, existing related technical solutions still have certain limitations in terms of data processing, user demand analysis, matching accuracy, and evaluation strategies, making it difficult to meet the needs of efficient and accurate housing replacement.

[0003] CN113903091B discloses a method, device, and storage medium for verifying the matching of houses and residents. This patent achieves accurate verification by setting parameters such as verification cycle, number of check-ins, and check-in time periods, and combining check-in information to determine the matching result between residents and houses. However, this technical solution mainly focuses on the verification and matching of residents and houses, lacking in-depth extraction and analysis of house feature information, and failing to fully consider the personalized needs of users and their dynamic changes in residential replacement. Furthermore, this solution does not provide a multi-dimensional evaluation mechanism for the matching results, which may lead to insufficient accuracy of the matching results and insufficient user satisfaction. CN118587461B discloses a model matching method, device, and storage medium for house layouts. This patent solves the problem of manual marking in matching house layouts and scanned data by dividing the house layout into multiple measurement stations and using point cloud data to generate the optimal rectangle for matching.

[0004] Because there is a lack of effective means to understand user needs, and because information obtained through surveys and other methods is difficult to quantify and accurately match, how to accurately obtain users' replacement needs has always been an urgent problem to be solved. The current general method is to analyze user input information, user access behavior, and user basic information to obtain user preferences. However, since housing information involves many dimensions, it is difficult to match through language description or behavioral analysis. Furthermore, the existing housing information is divided into low granularities, making it difficult to accurately determine the match between each house and user needs. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a residential replacement matching system based on information analysis.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is: a residential replacement matching system based on information analysis, including a screen feature processing subsystem, a user demand analysis module, a replacement matching index module, and a matching preference filtering module; The image feature processing subsystem is configured with a feature extraction strategy, which is used to acquire the user's VR image of the house to generate house feature information. The user demand parsing module is configured with a demand generation strategy, which is used to generate replacement demand information based on housing feature information and user information. The replacement matching index module is configured with a demand index strategy. The demand index strategy sets a corresponding index priority for each demand item in the replacement demand information, generates a matching index clue according to the demand item corresponding to the index priority, and obtains the corresponding house matching information from the house information database according to the matching index clue until the set of house matching information meets the matching conditions. The house matching information includes the house feature information. The matching preference filtering module is configured with a matching evaluation strategy. The matching evaluation strategy includes generating a matching requirement parameter group based on replacement requirement information and reorganizing user preference weights to construct a matching evaluation algorithm. The matching evaluation algorithm is used to calculate the corresponding matching evaluation value for each house matching information and output the matching result based on the matching evaluation value.

[0007] Firstly, by analyzing VR images on traditional platforms, a basis for understanding user needs is established. Based on the files uploaded by users to the housing platform, whether users are dissatisfied with their current housing or are looking to change for a specific purpose, VR images offer more dimensional possibilities in terms of information detail compared to information analysis and behavioral analysis. This makes user needs more concrete and specific. At the same time, this method allows for the analysis and quantification of housing listings in the same dimensions. Through demand analysis and replacement matching, information can be quickly indexed, improving data reliability and ensuring efficient matching of valid housing data. With image feature recognition technology as the core technology, a reliable basis for matching users with suitable housing is provided even with limited user information.

[0008] Furthermore, the feature extraction strategy includes: The image filtering step involves extracting a scene dataset from VR images according to preset scene extraction rules. The scene dataset satisfies preset feature abundance constraints and includes several house scene images. The image labeling step involves using a type matching algorithm to label the type of each house scene image. The feature extraction step is configured with a preset feature extraction sub-strategy. The feature extraction sub-strategy is used to filter feature map areas in the house image and to identify the feature information of each feature map area through a feature recognition model to obtain feature sub-labels.

[0009] By segmenting VR images into a set of images through image filtering, the volume, shape, and other features of objects can be extracted. At the same time, a type matching algorithm is used to label the images with their corresponding location and region type when identifying image features. Feature recognition is then performed based on type segmentation, which improves recognition efficiency and establishes more associations through the labeling relationship between regions and image features. When performing demand analysis or matching, the features have more recognition dimensions and can better reflect the needs between features and regions. Then, feature map areas are used to identify each piece of furniture, utensils, and decoration features in separate zones, improving recognition accuracy while ensuring that all information in the image is fed back and collected as much as possible.

[0010] Furthermore, the image extraction rules include acquiring continuous house images at preset acquisition intervals and acquisition start positions, with adjacent house images satisfying comprehensive constraints. These comprehensive constraints include overlap constraints, feature continuity constraints, and sharpness constraints. The overlap constraints reflect the degree of overlap between image images, the feature continuity constraints reflect the continuity of the same feature map region in the house image, and the sharpness constraints reflect the sharpness difference between adjacent images. The feature abundance constraints include calculating the feature abundance value of each house image to ensure that the mean feature abundance is greater than the feature abundance benchmark value.

[0011] By setting it up in this way, different VR images are processed into images based on the clarity, overlap, and feature continuity of the image extraction rules. This ensures that information is fully represented while keeping the amount of data small and reducing the consumption of computing resources. Through evaluation in three dimensions, the extraction rules for each user are different according to the characteristics of the house and the content, ensuring that user information is extracted as fully as possible.

[0012] Furthermore: the demand generation strategy includes: It is configured with several demand items, and the demand data corresponding to each demand item is matched from the house feature information based on the demand items; Each requirement item corresponds to a parsing sub-model, which is used to parse user information to obtain the requirement adjustment value of the corresponding requirement item; The corresponding demand data is updated based on the demand adjustment value to generate the replacement demand information.

[0013] The requirement generation strategy presents and aggregates user requirement characteristics by configuring requirement items. Different parsing sub-models are configured for different requirement items, which can break down the user's requirement for each dimension to quantify the user's requirement data for each dimension. Requirement data is generally presented in a vectorized way, with different vector components set to represent the user's requirement in a certain dimension. By using parsing sub-models, adjustments can be made to the original requirements based on the user's baseline requirements, combined with user information such as user input information, user basic information, and user behavior information.

[0014] Furthermore, the required items include storage requirements, living requirements, interference requirements, indoor requirements, environmental requirements, cost requirements, and location requirements. The storage requirements reflect the user's storage needs, the living requirements reflect the user's living area needs, the interference requirements reflect the user's public area needs, the indoor requirements reflect the user's indoor comfort needs, the environmental requirements reflect the user's living environment needs, the cost requirements reflect the user's replacement cost needs, and the location requirements reflect the user's living location needs.

[0015] By configuring different user needs for storage, living, interference, indoor, environmental, cost, and location, the system can first analyze user needs based on demand data. For example, storage needs can be categorized into different types, such as bathroom storage, laundry storage, and clothing storage, which can be further refined and quantified to more accurately reflect the user's overall storage requirements. The same applies to other aspects of needs. Living needs can include the number of bedrooms and the number of beds required. Interference needs include requirements for public areas such as balcony space, number of bathrooms, and dining areas. Indoor needs include factors such as decoration, lighting, ventilation, facility completeness, and yard size. Outdoor needs include requirements for hallway environment, lobby environment, public area environment, and property management. Cost needs include water and electricity bills, property management fees, and rent. Location needs include requirements for the surrounding environment and public facilities. This allows for a comprehensive and detailed analysis of user needs. The basic information for these needs is collected through public platforms based on VR images. Compared to traditional methods, it would be extremely difficult to collect such a large amount of information. Therefore, this comparison work is usually done collaboratively.

[0016] Furthermore: the parsing sub-model is configured with an information crawling library, which stores several information crawling clues and corresponding feature triples. The parsing sub-model extracts the triple structure features corresponding to the feature triples from the user information according to the matching relationship of the information crawling clues, and parses the triple structure features according to the user semantic preference library to obtain the corresponding demand adjustment value.

[0017] The information retrieval library configures information retrieval clues by constructing triples, allowing each different piece of information to be configured with a different construction method. For example, the triple for a wall is "material-texture-color". By abstractly extracting each feature, the features of different types of clues can be refined, while also satisfying the goal of quickly establishing different feature recognition methods. The pre-constructed feature triples ensure the efficiency of information retrieval and recognition, meet the standardization of features during information processing, and ensure the generation of required adjustment values.

[0018] Furthermore, the demand indexing strategy also includes a demand parsing sub-strategy, which is configured with a preference calculation algorithm. The preference calculation algorithm is used to calculate the user preference function for each demand item based on the historical user behavior in the user information, and generate the corresponding index priority based on the user preference function.

[0019] The demand indexing strategy includes a demand parsing sub-strategy. This sub-strategy uses a configuration preference calculation algorithm to parse demand, analyze the user preference function for each demand item, and generate a priority value corresponding to the index based on the user preference function. In this way, by analyzing historical user behavior, we can identify users' preferences for different demand items. For example, if users generally browse scenarios where renovation needs are preferred over their current living conditions, we can infer that users have a high preference for renovation. This preference can then be adjusted using the user preference function. By indexing using the user preference function, we can quickly identify and filter user needs based on a large amount of data.

[0020] Furthermore, the matching conditions are configured to have an abundance value based on the index priority of each demand item. When the types of housing matching information corresponding to each demand item all meet the abundance value requirements, the matching conditions are considered to be met. The abundance value reflects the degree of difference among all housing matching information under the demand item.

[0021] Because the selected housing information needs to be diverse to ensure that users have different criteria for selection, on the one hand, user needs can be further refined and updated through user selection behavior, and on the other hand, it can ensure that the houses identified by the algorithm can be selected and recognized by users as much as possible, saving users time to reselect.

[0022] Furthermore, the matching evaluation strategy includes matching the component sub-functions corresponding to each requirement item from a preset function component library based on user information, generating a benchmark sub-function for the requirement item based on the matching requirement parameters in the matching requirement parameter group, obtaining the requirement sub-function by subtracting the component sub-function and the benchmark sub-function, and constructing the matching evaluation algorithm by weighting the requirement sub-functions corresponding to each user preference weight in the user preference weight reorganization.

[0023] By setting up a matching evaluation strategy, it is ensured that user needs can be matched through component functions. For example, user needs for different needs are not linear, so according to the identification results, user preferences for specific houses are also not linear. Therefore, component functions can better reflect the user's specific preferences for a certain need. For example, the cost of rent may be acceptable to the user within a certain range, but if it exceeds this range, this need may prevent other needs from being selected even if they are matched. Therefore, by calling component functions, it is ensured that the user's information screening criteria are reasonable and reliable, while making the generated matching evaluation algorithm consistent with the user's actual need screening logic, so as to accurately determine the user's actual needs.

[0024] The main technical advantages of this invention are reflected in the following aspects: The system generates concrete house features from VR images of houses, combines them with user information to generate precise replacement needs; it efficiently filters house matching information according to the priority of the demand item index, and then calculates the evaluation value and outputs the result through an evaluation algorithm containing user preferences. This solves the problems of difficult demand acquisition and low matching granularity in traditional systems, improves matching efficiency and accuracy, ensures the reliability of housing matching, and better meets users' housing replacement needs. Attached Figure Description

[0025] Figure 1 This invention discloses a system architecture block diagram of a residential replacement matching system based on information analysis. Figure 2 : Flowchart of a feature extraction strategy for a residential replacement matching system based on information analysis according to the present invention; Figure 3 : A flowchart of a demand indexing strategy for a residential replacement matching system based on information analysis, as described in this invention; Figure 4 : A block diagram of a matching evaluation strategy for a residential replacement matching system based on information analysis according to the present invention; Figure 5 The present invention provides a flowchart for determining matching conditions in a residential replacement matching system based on information analysis. Detailed Implementation

[0026] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, so that the technical solution of the present invention can be more easily understood and mastered.

[0027] Reference Figures 1-5 As shown, a residential replacement matching system based on information analysis includes a visual feature processing subsystem, a user demand analysis module, a replacement matching index module, and a matching preference filtering module. Reference Figure 2As shown, the image feature processing subsystem is configured with a feature extraction strategy, which is used to acquire the user's VR image of the house to generate house feature information; The feature extraction strategy includes: The image filtering step extracts an image dataset from VR images according to preset image extraction rules. This image dataset satisfies preset feature abundance constraints and includes several house images. The image extraction rules involve acquiring consecutive house images at preset acquisition intervals and starting positions. Adjacent house images satisfy comprehensive constraints, including overlap, feature continuity, and sharpness constraints. The overlap constraint reflects the degree of overlap between images; the feature continuity constraint reflects the continuity of the same feature map area within the house images; and the sharpness constraint reflects the sharpness difference between adjacent images. The feature abundance constraint calculates the feature abundance value for each house image to ensure the mean feature abundance is greater than a baseline feature abundance value. The goal of the image filtering step is to extract an image dataset from the VR images that satisfies information integrity and validity. This step must strictly adhere to the preset image extraction rules and feature abundance constraints. The specific implementation is as follows: the image extraction rules are implemented by dynamically adjusting the acquisition interval and fixing the acquisition starting position. The starting position for image acquisition is uniformly set to 1 meter directly in front of the inside of the house's entrance door. The extraction direction is clockwise, traversing the entire house space. The acquisition interval is dynamically set according to the house's area: 0.5 seconds / frame for large spaces with an area ≥ 20㎡, and 1 second / frame for small spaces with an area < 20㎡. This setting reduces data redundancy while ensuring that no spatial features are missed. Comprehensive constraints include overlap constraints, feature continuity constraints, and sharpness constraints. All three must be satisfied simultaneously to retain the image frame. The specific calculation methods and constraint standards are as follows: Overlap Constraint: Overlap refers to the percentage of pixels in the overlapping area between two adjacent house images, used to balance information repetition and spatial continuity. Its calculation formula is: ,in Indicates the degree of overlap. This refers to the total number of pixels in the overlapping area where the pixel values ​​are completely identical in two adjacent frames of an image. This refers to the total number of pixels in the previous frame. The overlap constraint criterion in this system is... If the calculation result exceeds this range, the next frame image is discarded and re-extracted until the constraint is met. Feature continuity constraint: Feature continuity refers to the degree to which the same feature map area of ​​a house (such as the sofa in the living room, the bed in the bedroom, the cabinets in the kitchen, etc.) remains intact and its position is consistent in two adjacent frames. Its calculation formula is: ,in Indicates feature continuity, This refers to the number of complete pixels in the same feature map region that are not fragmented in two adjacent image frames. This refers to the total number of pixels in the feature map region within a single frame of an image. The constraint criterion for feature continuity in this system is... If the continuity of a certain feature map region is lower than a certain value, then the image frame is determined to be unable to fully reflect the feature information and needs to be re-extracted. Sharpness constraint: Sharpness is measured by the image grayscale gradient value and is used to avoid excessive fluctuations in sharpness between adjacent frames affecting feature recognition. First, the sharpness value of a single frame image is calculated using the Laplacian operator method. The specific method involves performing a Laplacian convolution operation on the image, taking the absolute value, and then calculating the average of all pixels as the G value for that frame. The formula for calculating the sharpness difference between two adjacent frames is: ,in This represents the resolution value of the current frame. This is the sharpness value of the previous frame. In this system, the constraint standard for sharpness difference is ΔG≤10. If it exceeds this range, the playback speed of the VR image is adjusted and the image is re-extracted. Implementation of feature abundance constraints: Feature abundance constraints are used to ensure that the image dataset covers the core feature elements of each space in the house, avoiding incomplete feature recognition due to insufficient information in the extracted images. First, the "feature abundance value" is defined as the ratio of the number of actually identified house feature elements in a single frame image to the total number of typical feature elements of that type of space. The calculation formula is: Where E represents the feature abundance value of a single frame image, This refers to the number of feature elements (such as sofas, coffee tables, beds, wardrobes, cabinets, windows, etc.) actually identified in the image frame. This refers to the total number of typical characteristic elements pre-defined for this type of space (a spatial characteristic database needs to be established in advance, such as the number of elements in the living room). Set at 10, including sofa, coffee table, TV, TV cabinet, lighting fixtures, wall decorations, windows, curtains, flooring, and electrical outlets; bedroom. Set to 8, including bed, wardrobe, bedside table, lamps, windows, curtains, flooring, and electrical outlets; kitchen... Set to 7, including cabinets, stove, sink, refrigerator space, lighting fixtures, flooring, and sockets. The mean feature abundance is the sum of all extracted frames. The arithmetic mean of the values ​​is calculated using the following formula: Where n is the total number of frames extracted from the image. to Frame 1 to Frame 2 The feature abundance value of the frame. The constraint standard for the mean feature abundance in this system is... ≥60%; if not, increase the number of image extraction frames for the corresponding space until... Once the standards are met, a qualified image dataset will be formed.

[0028] The image labeling step involves labeling each house image with its corresponding house space type using a type matching algorithm. The goal of this step is to label each frame in the image dataset with the corresponding house space type, enabling targeted feature extraction based on the space type. The specific implementation is as follows: Selection and training of the type matching algorithm: This system uses the ResNet50 convolutional neural network model as the core of the type matching algorithm. This model is a well-known image classification model in the field of computer vision, characterized by high recognition accuracy and high computational efficiency. The model training data uses the publicly available House3D house space image dataset, which contains over 100,000 house space images of different floor plans and decoration styles, labeled with six space types: "living room," "master bedroom," "secondary bedroom," "kitchen," "bathroom," and "balcony." During training, the dataset was divided into training, validation, and test sets in a 7:2:1 ratio. The cross-entropy loss function was used as the optimization objective, and stochastic gradient descent (SGD) was employed for model training. The number of training iterations was set to 50 rounds, and the initial learning rate was set to 0.001, which was decreased with each iteration. Ultimately, the model achieved a spatial type classification accuracy of ≥95% on the test set, ensuring the reliability of the labeling results.

[0029] The image labeling process is as follows: The image dataset obtained through the image filtering step is input into the trained ResNet50 model. The model outputs the spatial type probability corresponding to each frame, and the spatial type with the highest probability is taken as the labeling result for that frame. The labeling information is attached to the image file in the form of metadata, with the metadata format being "spatial type-extraction date-frame number". For example, "living room-20240520-001" represents the first frame of the living room image extracted on May 20, 2024, and "master bedroom-20240520-015" represents the 15th frame of the master bedroom image extracted on May 20, 2024. After labeling, 10% of the images are manually sampled. If the accuracy of the sampled images is consistent with the accuracy of the model's test set, the labeling is considered qualified; if there is a deviation, the model is readjusted and the images are labeled again.

[0030] The feature extraction step is configured with a preset feature extraction sub-strategy. This sub-strategy is used to filter feature regions in the house image and identify the feature information of each feature region through a feature recognition model to obtain feature sub-labels. The goal of the feature extraction step is to segment independent feature regions from images with labeled spatial types and identify detailed information of each feature region to form feature sub-labels. The feature extraction sub-strategy is divided into two stages: "region segmentation" and "feature recognition". The first stage is region segmentation, which uses the U-Net semantic segmentation model to process images with labeled spatial types. The model needs to be pre-trained for feature elements of different spatial types. For example, for a living room image, the model needs to be able to segment the regions corresponding to feature elements such as sofas, coffee tables, and televisions. The segmentation accuracy requirement is that the pixel deviation between the edge of the feature region and the edge of the actual feature element is ≤5 pixels. If the deviation exceeds the range, the segmentation threshold of the model is adjusted and re-segmented. Application of the feature recognition model: Feature recognition uses the YOLOv8 object detection model to identify information in each feature region obtained from region segmentation. The identified content includes three types of information: feature type, specification parameters, and attribute features. The specification parameters need to be calculated using the mapping relationship between image pixels and actual size. The mapping formula is as follows: ,in Indicates the actual size of the feature element. This represents the pixel length of the feature map region in the image. This indicates the actual distance between the camera and the feature element during VR image capture. This indicates the camera's focal length. Attribute features include color, material, etc., which are obtained through model analysis of the texture and color features of the feature map area, such as "gray fabric sofa" and "light brown solid wood coffee table".

[0031] Feature sub-tag generation: The recognition results of each feature map area are summarized into feature sub-tags. The format of the feature sub-tags is "space type-feature type-specification parameter-attribute feature", such as "living room-fabric sofa-length 2.2 meters-gray", "kitchen-solid wood cabinet-height 2.4 meters-white", and "bedroom-solid wood bed-width 1.8 meters-dark brown". All feature sub-tags are stored according to space type, ultimately forming complete house feature information. This information contains all feature element details of each space in the house and can be directly used for subsequent user requirement analysis.

[0032] The core function of the image feature processing subsystem is to transform user-uploaded VR images of houses into structured and quantifiable house feature information. This provides accurate and comprehensive data support for the subsequent user demand analysis module to generate replacement demand information and the replacement matching index module to filter house matching information. This subsystem completes the image processing, spatial type labeling, and feature element recognition of VR images step by step through a preset feature extraction strategy. The entire process must meet operability and repeatability requirements, ensuring that technicians in the relevant technical field can deploy and run the system based on the following content.

[0033] Reference Figure 3 As shown, the user demand parsing module is configured with a demand generation strategy, which is used to generate replacement demand information based on house feature information and user information. The core function of the user demand parsing module is to receive the house feature information output by the image feature processing subsystem and combine it with user information to generate structured and quantifiable replacement demand information. At the same time, it determines the index priority of each demand item through the demand parsing sub-strategy, providing accurate demand basis for the subsequent replacement matching index module.

[0034] The demand generation strategy includes: The system is configured with several requirement items. For each requirement item, the corresponding requirement data is matched from the house feature information. Requirement data matching involves extracting baseline data corresponding to each requirement item's sub-dimension from the house feature information output by the image feature processing subsystem, forming an initial requirement dataset. The specific matching process is as follows: First, a mapping table between requirement item sub-dimensions and house feature information is established. For example, "Storage Requirement Item - Clothing Storage Sub-dimension" is mapped to the volume parameters of all feature sub-tags such as "Bedroom - Wardrobe" and "Bedroom - Dresser" in the house feature information. Then, baseline values ​​for each sub-dimension are obtained through data statistics and calculations. Volume parameters are calculated using summation, quantity parameters using counting, area parameters using direct extraction, and attribute parameters using categorized summation.

[0035] Each requirement item corresponds to a parsing sub-model, which is used to parse user information to obtain the requirement adjustment value of the corresponding requirement item; The parsing sub-model is configured with an information retrieval library, which stores several information retrieval clues and corresponding feature triples. The parsing sub-model extracts the triple structure features corresponding to the feature triples from user information based on the matching relationships of the information retrieval clues, and parses the triple structure features according to a user semantic preference library to obtain the corresponding demand adjustment value. The core of the parsing sub-model is to extract key information affecting demand from user information and calculate the demand adjustment value to correct the baseline demand data. Its execution steps are as follows: The construction process of the information retrieval library is as follows: The information retrieval library needs to pre-store information retrieval clues and feature triples corresponding to the seven major demand items and sub-dimensions. Information retrieval clues are keywords or behavioral identifiers that identify demand-related information from user information. For example, the retrieval clues for the storage demand item include identification features such as "bedroom," "cabinet," "handle," and "open," which makes it easier for the system to lock the type for feature extraction during image processing. Feature triples are three sets of related information used to structurally describe the key attributes of the demand item. The format is "demand dimension-attribute type-current state." For example, the feature triple for the storage demand item can be set as "clothing storage-space capacity-1.2 cubic meters"; in addition, if the wardrobe is in the closed state, then it is "clothing storage-projection capacity-1 square meter."

[0036] Feature triple extraction and parsing: Based on the matched information crawling clues, the corresponding feature triples are retrieved from the information crawling library, and then combined with the user semantic preference library to parse the triple structure features to determine the demand expectation. The user semantic preference library is a pre-built database containing semantic mapping relationships. For example, "increase storage space" is mapped to "demand capacity is 30% higher than the current capacity", and "near subway" is mapped to "commuting time ≤ 20 minutes".

[0037] Calculation of demand adjustment value: The demand adjustment value adopts a linear calculation model, and the formula is as follows: ,in This represents the adjustment value of a sub-dimension of a requirement item. Indicates the adjustment factor. This represents the expected demand obtained from parsing the user's semantic preference library. This represents the baseline demand data matched from housing feature information. Adjustment factor. The level of attention to this need is set based on the number of times users mention it or the frequency of their actions. For example, if a user mentions "increase clothing storage" three times, then... Take 1.2, if it is mentioned only once. If α is set to 1.0, and if no relevant browsing behavior is mentioned, then α is set to 0.8. The value range is 0.8-1.2 to ensure that the adjustment range is reasonable.

[0038] The required items include storage requirements, living requirements, interference requirements, indoor requirements, environmental requirements, cost requirements, and location requirements. The storage requirements reflect the user's storage needs, the living requirements reflect the user's residential area requirements, the interference requirements reflect the user's public area requirements, the indoor requirements reflect the user's indoor comfort requirements, the environmental requirements reflect the user's residential environment requirements, the cost requirements reflect the user's replacement cost requirements, and the location requirements reflect the user's residential location requirements.

[0039] The demand generation strategy first requires breaking down the seven pre-defined demand items into sub-dimensions. Each sub-dimension of a demand item must correspond one-to-one with the feature sub-tags in the housing feature information and the associated dimensions of the user information to ensure accurate matching of demand data. The specific details are as follows: Storage requirements include sub-dimensions for bathroom storage, laundry and drying storage, and clothing storage. The bathroom storage sub-dimension corresponds to features such as "Bathroom - Shelf - Capacity 0.3 cubic meters" and "Bathroom - Mirror Cabinet - Capacity 0.2 cubic meters" in the house feature information. The laundry and drying storage sub-dimension corresponds to features such as "Balcony - Laundry Cabinet - Capacity 0.5 cubic meters" and "Balcony - Clothes Rack - Load Capacity 5 kg" in the house feature information. The clothing storage sub-dimension corresponds to features such as "Bedroom - Wardrobe - Volume 1.2 cubic meters" and "Bedroom - Dresser - Volume 0.4 cubic meters" in the dresser.

[0040] Housing requirements include the following sub-dimensions: number of bedrooms, bed size, and occupant matching. The number of bedrooms corresponds to the number of space type tags for "master bedroom" and "secondary bedroom" in the house feature information. The bed size corresponds to feature tags such as "bedroom - solid wood bed - 1.8 meters wide". The occupant matching sub-dimension is determined based on the number of family members in the user information. For example, a family of 3 people needs at least 2 bedrooms.

[0041] Interference requirements include the public area area sub-dimension, the number of toilets sub-dimension, and the dining space sub-dimension. The public area area sub-dimension is calculated from the area of ​​"living room" and "balcony" in the house feature information. The number of toilets sub-dimension corresponds to the number of space type markers for "toilet". The dining space sub-dimension corresponds to feature sub-markers such as "living room - dining table - area of ​​0.8 square meters".

[0042] Indoor requirements include the following sub-dimensions: lighting, ventilation, decoration style, and facility completeness. The lighting sub-dimension corresponds to features such as "living room - window - area 1.5 square meters" and "bedroom - window - south-facing". The ventilation sub-dimension corresponds to the number of windows and their opening methods, such as "living room - window - sliding". The decoration style sub-dimension corresponds to the attribute characteristics of features such as "living room - wall - latex paint - white" and "floor - solid wood - light brown". The facility completeness sub-dimension corresponds to features such as "kitchen - gas stove - brand XX" and "bathroom - water heater - type electric water heater".

[0043] Environmental requirements include the following sub-dimensions: corridor environment, public area environment, and property services. The corridor environment sub-dimension is extracted from user information such as "mentioning corridor cleanliness". The public area environment sub-dimension corresponds to user information such as "hoping for greenery in the community". The property services sub-dimension corresponds to user browsing history and the behavior of clicking on "24-hour security" listings.

[0044] Cost requirements include sub-dimensions for utilities and property management fees, rent, and replacement fees. The utilities and property management fees sub-dimension is extracted from the user's input text "monthly utilities and property management fees not exceeding 500 yuan." The rent sub-dimension corresponds to the rent range of properties the user has historically viewed (e.g., 1500-2000 yuan / month). The replacement fees sub-dimension corresponds to the user's statement "desiring the fees to be less than 1% of the total house price." Alternatively, the user's cost requirements can be extracted from the current property information.

[0045] Location requirements include the following sub-dimensions: commuting distance, surrounding amenities, and regional preference. The commuting distance sub-dimension is extracted from the user information, which states that "commuting to A1 company takes no more than 30 minutes." The surrounding amenities sub-dimension corresponds to text such as "hoping for supermarkets and schools nearby." The regional preference sub-dimension corresponds to the user's historical browsing history of properties in the "A3 area."

[0046] It should be noted that there can be multiple ternary feature groups for each item or feature. For example, they can represent color, material, or position, without limitation.

[0047] The baseline demand data for each sub-dimension of demand items are added to the demand adjustment value to obtain the corrected demand target value. All demand target values ​​are summarized by demand item to form the final replacement demand information.

[0048] The demand parsing sub-strategy calculates the user's preference for each demand item using a preference calculation algorithm, generating index priorities to guide subsequent replacement matching. The core of the preference calculation algorithm is constructing a user preference function, the formula of which is... ,in Indicates the user's opinion on the first The preference value of each demand item. Indicates the weight of the number of interactions. Indicates the weight of dwell time. This represents the number of times a user interacts with information related to the i-th request item. Indicates that the user browsed the first The total dwell time on information related to each demand item This represents the normalized coefficient representing the number of interactions and dwell time for all request items. , This represents the total number of interactions for all requirement items. This indicates the total dwell time for all demand items. and The value is always 0.5 to ensure that the number of interactions and the duration of the interaction are equally important, preference value The value ranges from 0 to 1; a larger value indicates a higher degree of user preference for that requirement. Sort the items from largest to smallest to obtain their index priority.

[0049] Reference Figure 3 As shown, the replacement matching index module is configured with a demand index strategy. The demand index strategy sets a corresponding index priority for each demand item in the replacement demand information, and generates matching index clues according to the demand items corresponding to the index priorities. It then matches the housing information database according to the matching index clues to obtain the corresponding housing matching information until the set of housing matching information meets the matching conditions. The housing matching information includes the housing feature information. The demand index strategy also includes a demand parsing sub-strategy, which is configured with a preference calculation algorithm. The preference calculation algorithm is used to calculate the user preference function for each demand item based on the historical user behavior in the user information, and generates the corresponding index priority based on the user preference function.

[0050] The core function of the replacement matching index module is to take the replacement demand information and index priority of each demand item output by the user demand parsing module, generate matching index clues through the preset demand index strategy, retrieve and filter house matching information from the house information database until the set of house matching information corresponding to all demand items meets the matching conditions, and provide candidate house data that meet the requirements for the subsequent matching preference filtering module. The house matching information includes house feature information generated by the image feature processing subsystem.

[0051] Reference Figure 5As shown, matching index clues refer to the set of filtering conditions generated based on the target value and index priority of each requirement item, used to accurately retrieve housing listings in the housing information database. They need to correspond one-to-one with the fields in the housing information database (such as commuting time, storage space, property fees, etc.). The specific steps of the demand indexing strategy are as follows: First, determine the priority of demand items. The preference values ​​output by the user demand parsing module are used for sorting. Demand items with higher preference values ​​have higher index priority. During retrieval, priority is given to filtering based on high-priority demand items to reduce invalid search data. Second, generate matching index clues. Filtering conditions are constructed for each demand item's sub-dimensional target value, ranked from highest to lowest priority. For example, if the highest priority location demand item's target value is "commuting time ≤ 22 minutes and there is a primary school within 3 kilometers," then the corresponding matching index clue is "commuting time field ≤ 22 minutes AND surrounding facilities field includes a primary school." Clues for lower-priority demand items need further filtering based on the search results of higher-priority clues to avoid redundant cross-priority searches. Third, execute database retrieval. The generated matching index clues are sequentially input into the housing information database. Each round of retrieval only retains housing data that matches the current clue, serving as the base data pool for the next round of low-priority clue retrieval. Fourth, determine if the matching conditions are met. Matching conditions need to be verified by calculating the abundance value of each demand item. The abundance value is an indicator reflecting the difference in matching information among all housing items under a certain demand item. The calculation formula is... ,in Indicates the first Abundance values ​​of each demand item This indicates the number of housing matching information types with differentiated characteristics under this demand item. This indicates the total number of housing matching information corresponding to this demand item in the current data pool. This system sets the abundance value of all demand items to be ≥30% to be considered as meeting the matching conditions. If the abundance value of a certain demand item does not meet the standard, it is necessary to supplement the search for other housing listings corresponding to the demand item clues until the abundance value meets the requirements. At this time, the search stops and all housing matching information in the current data pool is output.

[0052] The matching conditions are configured with an abundance value based on the index priority of each demand item. When the types of housing matching information corresponding to each demand item meet the abundance value requirements, the matching conditions are considered to be met. The abundance value reflects the degree of difference among all housing matching information under the demand item.

[0053] The matching preference filtering module is configured with a matching evaluation strategy. This strategy includes generating a matching requirement parameter set based on replacement requirement information and reorganizing user preference weights to construct a matching evaluation algorithm. The matching evaluation algorithm calculates the matching evaluation value corresponding to each house matching information and outputs the matching result based on the evaluation value. The matching evaluation strategy also includes matching the component sub-function corresponding to each requirement item from a preset function component library based on user information, generating a benchmark sub-function for that requirement item based on the matching requirement parameters in the matching requirement parameter set, subtracting the component sub-function from the benchmark sub-function to obtain the requirement sub-function, and weighting the corresponding requirement sub-function according to each user preference weight in the user preference weight reorganization to construct the matching evaluation algorithm.

[0054] The core function of the matching preference filtering module is to receive the housing matching information output by the replacement matching index module, combine it with the replacement demand information generated by the user demand analysis module, construct a matching evaluation algorithm through a preset matching evaluation strategy, calculate the matching evaluation value of each housing matching information, and finally output the matching results according to the evaluation value, providing users with candidate housing that meet their personalized preferences. The whole process must ensure that the algorithm logic is reproducible, the calculation steps are operable, and meet the implementation needs of the technical personnel in the relevant technical field.

[0055] The matching requirement parameter set refers to the set of target values ​​for each sub-dimension of the requirement items extracted from the replacement requirement information. Each requirement item corresponds to at least one parameter. The user preference weighting set refers to the set of weights proportionally allocated to the preference values ​​of each requirement item output by the user requirement analysis module. The total weight is 1, used to reflect the importance of different requirement items in the user's decision-making. For example, the user's preference value for the location requirement item is the highest, so its weight can be set to 0.3; the storage requirement item is next, set to 0.25; and the environment and cost requirements items are set to 0.2 and 0.25 respectively. The function composition library refers to a pre-set database storing the component sub-functions corresponding to each requirement item. The component sub-function is a mathematical function reflecting the user's actual preference for a certain requirement item. It needs to be matched according to the requirement item type. For example, the cost requirement item corresponds to a piecewise function, and the environment requirement item corresponds to a classification assignment function. Technical personnel can pre-configure functions based on common user preference types, or iteratively optimize the function form through user historical behavior data.

[0056] The first step is to extract the matching requirement parameter set: From the replacement requirement information, the target values ​​of each requirement item's sub-dimensions are filtered one by one, eliminating duplicate or redundant parameters to form a structured parameter set. For example, user B's comprehensive analysis yields the following requirement items: , , , , , , These correspond to storage, living, interference, indoor, environment, cost, and location.

[0057] The second step is to generate a restructured user preference weight: based on the preference values ​​of each requirement item output by the user requirement analysis module, the weights are calculated using normalization, as shown in the formula below. ,in Indicates the first The weight of each requirement item, Indicates the first The preference value of each demand item. This represents the sum of all preference values ​​for all demand items. For example, the preference value for user B's location demand item. , , , , , , .

[0058] The third step involves obtaining each component sub-function based on user information. These component sub-functions are obtained through user behavior. Then, a corresponding baseline sub-function is generated based on the user demand vector, thus constructing the matching evaluation algorithm. + + + + + + ,in, For the requirement item The corresponding storage sub-function, This is the storage demand vector deviation value, reflecting the discrepancy between actual storage demand and the corresponding available housing. For the requirement item The corresponding residential sub-function, This is the housing demand vector deviation value, reflecting the discrepancy between actual housing demand and the corresponding housing supply. For the requirement item The corresponding interferometer function, This is the interference vector deviation value, reflecting the discrepancy between the actual interference and the corresponding housing demand. For the requirement item The corresponding indoor sub-function, This is the indoor demand vector deviation value, reflecting the deviation between actual indoor demand and the demand for corresponding housing units. For the requirement item The corresponding environment subfunction, This is the environmental demand vector deviation value, reflecting the deviation between actual environmental demand and the corresponding housing demand. For the requirement item The corresponding cost subfunction, This is the cost demand vector deviation value, reflecting the discrepancy between actual cost demand and the corresponding housing demand. For the requirement item The corresponding positional subfunction, This is the location demand vector deviation value, reflecting the deviation between the actual location demand and the corresponding housing demand. By setting it in this way, the deviation can be quantified after each housing information is set, and the matching results can be calculated. The results are then sorted according to the result value corresponding to each housing, and the information comparison output is realized.

[0059] Of course, the above are just typical examples of the present invention. In addition, the present invention may have many other specific embodiments. All technical solutions formed by equivalent substitution or equivalent transformation fall within the scope of protection claimed by the present invention.

Claims

1. A residential replacement matching system based on information analysis, characterized in that: It includes a screen feature processing subsystem, a user demand analysis module, a replacement matching index module, and a matching preference filtering module; The image feature processing subsystem is configured with a feature extraction strategy, which is used to acquire the user's VR image of the house to generate house feature information. The user demand parsing module is configured with a demand generation strategy, which is used to generate replacement demand information based on housing feature information and user information. The replacement matching index module is configured with a demand index strategy. The demand index strategy sets a corresponding index priority for each demand item in the replacement demand information, and generates a matching index clue according to the demand item corresponding to the index priority. The matching index clue is used to match the corresponding house matching information from the house information database until the set of house matching information meets the matching conditions. The house matching information includes the house feature information. The matching preference filtering module is configured with a matching evaluation strategy. The matching evaluation strategy includes generating a matching requirement parameter group based on replacement requirement information and reorganizing user preference weights to construct a matching evaluation algorithm. The matching evaluation algorithm is used to calculate the corresponding matching evaluation value for each house matching information and output the matching result based on the matching evaluation value.

2. The residential replacement matching system based on information analysis as described in claim 1, characterized in that: The feature extraction strategy includes: The image filtering step involves extracting a scene dataset from VR images according to preset scene extraction rules. The scene dataset satisfies preset feature abundance constraints and includes several house scene images. The image labeling step involves using a type matching algorithm to label the type of each house scene image. The feature extraction step is configured with a preset feature extraction sub-strategy. The feature extraction sub-strategy is used to filter feature map areas in the house image and to identify the feature information of each feature map area through a feature recognition model to obtain feature sub-labels.

3. The residential replacement matching system based on information analysis as described in claim 2, characterized in that: The image extraction rules include acquiring continuous house images at preset acquisition intervals and acquisition start positions, with adjacent house images satisfying comprehensive constraints. These comprehensive constraints include overlap constraints, feature continuity constraints, and sharpness constraints. The overlap constraints reflect the degree of overlap between image images, the feature continuity constraints reflect the continuity of the same feature map region in the house image, and the sharpness constraints reflect the sharpness difference between adjacent images. The feature abundance constraints include calculating the feature abundance value of each house image to ensure that the mean feature abundance is greater than the feature abundance benchmark value.

4. The residential replacement matching system based on information analysis as described in claim 2, characterized in that: The demand generation strategy includes: It is configured with several demand items, and the demand data corresponding to each demand item is matched from the house feature information based on the demand items; Each requirement item has a corresponding parsing sub-model, which is used to parse user information to obtain the requirement adjustment value of the corresponding requirement item; The corresponding demand data is updated based on the demand adjustment value to generate the replacement demand information.

5. A residential replacement matching system based on information analysis as described in claim 4, characterized in that: The required items include storage requirements, living requirements, interference requirements, indoor requirements, environmental requirements, cost requirements, and location requirements. The storage requirements reflect the user's storage needs, the living requirements reflect the user's residential area requirements, the interference requirements reflect the user's public area requirements, the indoor requirements reflect the user's indoor comfort requirements, the environmental requirements reflect the user's residential environment requirements, the cost requirements reflect the user's replacement cost requirements, and the location requirements reflect the user's residential location requirements.

6. The residential replacement matching system based on information analysis as described in claim 4, characterized in that: The parsing sub-model is configured with an information crawling library, which stores several information crawling clues and corresponding feature triples. The parsing sub-model extracts the triple structure features corresponding to the feature triples from the user information based on the matching relationship of the information crawling clues, and parses the triple structure features based on the user semantic preference library to obtain the corresponding demand adjustment value.

7. A residential replacement matching system based on information analysis as described in claim 6, characterized in that: The demand indexing strategy also includes a demand parsing sub-strategy, which is configured with a preference calculation algorithm. The preference calculation algorithm is used to calculate the user preference function for each demand item based on the historical user behavior in the user information, and generate the corresponding index priority based on the user preference function.

8. A residential replacement matching system based on information analysis as described in claim 1, characterized in that: The matching conditions are configured with an abundance value based on the index priority of each demand item. When the types of housing matching information corresponding to each demand item meet the abundance value requirements, the matching conditions are considered to be met. The abundance value reflects the degree of difference among all housing matching information under the demand item.

9. A residential replacement matching system based on information analysis as described in claim 1, characterized in that: The matching evaluation strategy includes matching the component sub-functions corresponding to each requirement item from a preset function component library based on user information, generating a benchmark sub-function for the requirement item based on the matching requirement parameters in the matching requirement parameter group, obtaining the requirement sub-function by subtracting the component sub-function and the benchmark sub-function, and constructing the matching evaluation algorithm by weighting the requirement sub-functions corresponding to each user preference weight in the user preference weight reorganization.

Images