Railway tunnel operation and maintenance decision recommendation method and system based on collaborative filtering algorithm

Abstract

The invention discloses a railway tunnel operation and maintenance decision recommendation method based on a collaborative filtering algorithm, and the method comprises the following steps: obtaining related data of a plurality of railway tunnels, and carrying out the preprocessing; the related data comprises basic information, geographic position information and historical maintenance information; key feature extraction is carried out on the preprocessed related data, and a user-operation and maintenance task matrix is constructed; based on the user-operation and maintenance task matrix, the similarity between users and the similarity between railway tunnel operation and maintenance tasks are calculated through Pearson's correlation coefficients; and based on the similarity between the users and the similarity between the railway tunnel operation and maintenance tasks, adopting a collaborative filtering algorithm to generate an operation and maintenance decision corresponding to the maintenance request. According to the method, the efficiency and accuracy of railway tunnel operation and maintenance work are improved, the maintenance work is promoted to be changed from passive response to active prevention, the predictive processing capacity of faults is enhanced, and then the operation safety and reliability of the railway tunnel are comprehensively enhanced.

Description

Technical Field

[0001] This invention relates to the field of railway tunnel operation and maintenance management, and more specifically to a railway tunnel operation and maintenance decision recommendation method and system based on collaborative filtering algorithm. Background Technology

[0002] In the field of railway tunnel operation and maintenance management, the continuous expansion of railway networks and the increase in the number of tunnels have made efficient and accurate operation and maintenance decisions a significant challenge for the industry. Traditional operation and maintenance management methods often rely on experience-based judgment or periodic inspections, which are not only inefficient but also ill-suited to handling unexpected situations and complex and ever-changing operation and maintenance needs.

[0003] Traditional operation and maintenance (O&M) decision-making often relies on fragmented and disorganized information, which is frequently outdated, leading to insufficient decision-making basis and a high risk of errors or delays. For example, crucial information such as the tunnel's geographical location, climate conditions, and historical maintenance records may not be fully utilized, affecting the accuracy and timeliness of decisions. Furthermore, due to reliance on experience-based judgment, O&M decisions are often influenced by the personal experience and subjective opinions of management personnel, resulting in less objective and scientific outcomes. This highly subjective decision-making approach is particularly prone to disputes and errors when facing complex and ever-changing O&M problems.

[0004] Furthermore, different tunnels have varying operation and maintenance needs due to differences in geographical location, structural type, and operational status. However, existing technologies often adopt a "one-size-fits-all" approach to operation and maintenance management, lacking accurate identification and recommendations for the personalized operation and maintenance needs of specific tunnels. This leads to unreasonable allocation of operation and maintenance resources and low operation and maintenance efficiency.

[0005] Traditional operation and maintenance management systems struggle to handle the demands of large-scale data processing and analysis, exhibiting poor scalability and real-time performance. Furthermore, the lack of a dynamic update mechanism prevents the system from promptly reflecting the latest operational status and changes, impacting the effectiveness and timeliness of operational decisions.

[0006] Therefore, how to design a railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm to improve the accuracy and timeliness of operation and maintenance decisions and realize the optimal allocation and efficient utilization of operation and maintenance resources is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0007] In view of this, the present invention provides a railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm, which effectively integrates basic information, geographical data and historical maintenance records of railway tunnels, constructs a user-operation and maintenance task matrix, and realizes intelligent operation and maintenance decision recommendation by calculating the similarity between users and operation and maintenance tasks.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] In a first aspect, the present invention provides a railway tunnel operation and maintenance decision recommendation method based on a collaborative filtering algorithm, comprising the following steps:

[0010] S1. Acquire relevant data from multiple railway tunnels and perform preprocessing; the relevant data includes: basic information, geographical location information, and historical maintenance information;

[0011] S2. Extract key features from the preprocessed data and construct a user-operation task matrix;

[0012] S3. Based on the user-maintenance task matrix, calculate the similarity between users and between railway tunnel maintenance tasks using the Pearson correlation coefficient;

[0013] S4. Based on the similarity between users and between railway tunnel maintenance tasks, a collaborative filtering algorithm is used to generate maintenance decisions corresponding to maintenance requests.

[0014] Furthermore, in S2, the key features include:

[0015] The basic information includes the tunnel structure type, length, and construction year; the geographical location information includes longitude, latitude, altitude, and climate zone information; and the historical maintenance information includes the fault type, fault frequency, maintenance measures, and maintenance costs.

[0016] Furthermore, in S2, the user-operation and maintenance task matrix is ​​constructed, including:

[0017] S21. Define a user set U = {U1, U2, ..., U...} n} and the set of operation and maintenance tasks T = {T1, T2, ..., T m}; where each user U i This represents a tunnel management or maintenance team, with each maintenance task (T)... j This indicates a specific operational or maintenance requirement or project.

[0018] S22. Based on the user set U and the operation and maintenance task set T, construct the user-operation and maintenance task matrix R; where the elements R in the user-operation and maintenance task matrix R are... ij Indicates user U i For maintenance task T j Participation level;

[0019] S23. Combine the time decay function δ(t) and the weight coefficient ω to adjust the user-maintenance task matrix R, and obtain the adjusted user-maintenance task matrix R′.

[0020] Furthermore, in S23, the element R′ in the user-maintenance task matrix R′ ij Represented as:

[0021] R′ ij =δ(t)*ω ij *R ij

[0022] Wherein, the decay function δ(t) represents the user behavior weight at different times, δ(t) = e -λt λ represents the decay rate; weighting coefficient ω ij Indicates user U i During the execution of maintenance task T j Weighting different types of user behavior.

[0023] Furthermore, in S2, the user-operation task matrix is ​​stored using a sparse matrix storage format and incorporates a dynamic update mechanism; the dynamic update mechanism includes:

[0024] Acquire real-time data streams or periodically query new user behavior data;

[0025] Based on the user ID and operation and maintenance task ID in the new user behavior data, determine the affected elements in the user-operation and maintenance task matrix;

[0026] The affected elements are updated using an update function to obtain the updated user-operation task matrix.

[0027] Furthermore, in S3, calculating the similarity between users includes:

[0028] S311. Define a user vector for each user; the vector includes the user's participation in all operation and maintenance tasks;

[0029] S312. Calculate the covariance and standard deviation based on the user vectors of user A and user B;

[0030] The covariance Cov(A,B) is expressed as:

[0031]

[0032] The standard deviation includes σA and σB;

[0033]

[0034] Among them, A o B o These represent the degree of participation of user A and user B in the operation and maintenance task o, respectively. represents the average engagement level of user A and user B respectively, and m represents the number of maintenance tasks shared by user A and user B;

[0035] S313. Based on the covariance and standard deviation, determine the similarity R between users using the Pearson correlation coefficient. AB :

[0036]

[0037] Among them, R AB ∈[-1,1], where 1 represents a perfect positive correlation, -1 represents a perfect negative correlation, and 0 represents no linear correlation.

[0038] Furthermore, in S3, calculating the similarity between railway tunnel operation and maintenance tasks includes:

[0039] S321. Define an operation and maintenance task vector for each railway tunnel operation and maintenance task; the vector includes the participation degree of all users in the operation and maintenance task;

[0040] S322, Based on Operation and Maintenance Task A ′ Calculate the covariance and standard deviation of the operation and maintenance task vector and the operation and maintenance task B′.

[0041] The covariance Cov(A′,B′) is expressed as:

[0042]

[0043] The standard deviation includes σA ' and σB′;

[0044]

[0045]

[0046] Among them, A′ p B′ p Let A and B represent the degree of participation of user p in operation and maintenance task A' and operation and maintenance task B', respectively. These represent the user's views on maintenance task A' and maintenance task B, respectively. ' Average participation, l represents the average participation of operation and maintenance task A′ and operation and maintenance task B. ' Total number of users;

[0047] S323. Based on the covariance and standard deviation, determine the similarity R between users using the Pearson correlation coefficient. A′B′ :

[0048]

[0049] Among them, R A′B′ ∈[-1,1], where 1 represents a perfect positive correlation, -1 represents a perfect negative correlation, and 0 represents no linear correlation.

[0050] Furthermore, in step S4, a collaborative filtering algorithm is used to generate operation and maintenance decisions corresponding to the maintenance request, including:

[0051] Determining operational decisions based on user-side collaborative filtering:

[0052] S411. For each user u who needs to generate operation and maintenance decisions, determine the k users most similar to user u.

[0053] N u =argmax v≠u sim(u,v)

[0054] S412. Based on the historical participation of k users in the operation and maintenance task, generate a rating for user u on the operation and maintenance task w.

[0055]

[0056] Where sim(u,v) represents the similarity between user u and user v, r v,w This indicates the degree of participation of user v in the operation and maintenance task w;

[0057] S413. Determine the operation and maintenance decision based on the score.

[0058] Furthermore, in step S4, the process of generating operation and maintenance decisions corresponding to maintenance requests using a collaborative filtering algorithm also includes:

[0059] S421. Determine operation and maintenance decisions based on collaborative filtering on the operation and maintenance task side:

[0060] For each operation and maintenance task u′ to be decided, identify k′ operation and maintenance tasks that are most similar to operation and maintenance task u′;

[0061] N u′ =argmax v′≠u′ sim(u′,v′)

[0062] S422. Based on the user's historical participation in k′ operation and maintenance tasks, generate a rating for user w′ on operation and maintenance task u′;

[0063]

[0064] Where sim(u′,v′) represents the similarity between operation and maintenance task u′ and operation and maintenance task v′, r v′,w′ This represents the degree of participation of user w′ in operation and maintenance task v′;

[0065] S423. Determine the operation and maintenance decision based on the score.

[0066] Secondly, this invention provides a railway tunnel operation and maintenance decision recommendation system based on a collaborative filtering algorithm, comprising:

[0067] Data acquisition and preprocessing module: used to acquire relevant data from multiple railway tunnels and perform preprocessing; the relevant data includes: basic information, geographical location information, and historical maintenance information;

[0068] Feature extraction and matrix construction module: used to extract key features from preprocessed relevant data and construct a user-operation task matrix;

[0069] Similarity calculation module: used to calculate the similarity between users and between railway tunnel maintenance tasks based on the user-maintenance task matrix and the Pearson correlation coefficient;

[0070] Decision generation module: Based on the similarity between users and between railway tunnel maintenance tasks, it uses a collaborative filtering algorithm to generate maintenance decisions corresponding to maintenance requests.

[0071] As can be seen from the above technical solution, compared with the prior art, the technical solution of the present invention has the following beneficial effects:

[0072] 1. This method extracts multi-dimensional key features of railway tunnels and constructs a user-maintenance task matrix, achieving a precise characterization of maintenance needs. By combining Pearson correlation coefficient to calculate the similarity between users and maintenance tasks, the collaborative filtering algorithm can provide tunnel managers with highly personalized maintenance decision recommendations based on historical data and similar behavioral patterns, significantly improving maintenance efficiency and effectiveness.

[0073] 2. A sparse matrix storage format is adopted to store the user-operation task matrix, and a dynamic update mechanism is introduced. By capturing user behavior data in real time or periodically, the elements in the matrix are dynamically adjusted to ensure that the operation and maintenance decision recommendation system can keep up with changes in the actual situation and reflect the latest operation and maintenance needs and patterns in a timely manner. This strong real-time feature makes operation and maintenance decisions more realistic and reduces decision-making errors caused by information lag.

[0074] 3. Unlike traditional one-way collaborative filtering methods, this approach combines two-way collaborative filtering from both the user side and the operation and maintenance task side. User-side collaborative filtering predicts the current user's operation and maintenance needs based on the historical behavior of similar users; while operation and maintenance task-side collaborative filtering recommends suitable operation and maintenance measures to users based on historical data of similar operation and maintenance tasks. This two-way combination considers both the similarity between users and the correlation between operation and maintenance tasks, making operation and maintenance decisions more comprehensive and accurate, and improving the scientific nature and effectiveness of overall operation and maintenance management. Attached Figure Description

[0075] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0076] Figure 1 A flowchart of a railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm is provided in an embodiment of the present invention;

[0077] Figure 2 This is a framework diagram of a railway tunnel operation and maintenance decision recommendation system based on a collaborative filtering algorithm, provided for an embodiment of the present invention. Detailed Implementation

[0078] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0079] Example 1;

[0080] like Figure 1 As shown, this embodiment provides a railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm, including the following steps:

[0081] S1. Acquire relevant data from multiple railway tunnels and perform preprocessing; the relevant data includes: basic information, geographical location information, and historical maintenance information;

[0082] S2. Extract key features from the preprocessed data and construct a user-operation task matrix;

[0083] S3. Based on the user-maintenance task matrix, calculate the similarity between users and between railway tunnel maintenance tasks using the Pearson correlation coefficient;

[0084] S4. Based on the similarity between users and between railway tunnel maintenance tasks, a collaborative filtering algorithm is used to generate maintenance decisions corresponding to maintenance requests.

[0085] This method effectively integrates basic information, geographic data, and historical maintenance records of railway tunnels using a collaborative filtering algorithm to construct a user-maintenance task matrix. By calculating the similarity between users and maintenance tasks, it achieves intelligent maintenance decision recommendations. This not only improves the efficiency and accuracy of railway tunnel maintenance work but also promotes a proactive shift in maintenance efforts, thereby facilitating preventative maintenance, optimizing resource allocation, and ultimately enhancing the safety and reliability of railway tunnel operations.

[0086] The following provides a further detailed explanation of each step in the above method:

[0087] In this embodiment S1, relevant data of multiple railway tunnels are acquired and preprocessed; the relevant data includes: basic information, geographical location information, and historical maintenance information;

[0088] Specifically, preprocessing includes: data cleaning of the relevant data; the data cleaning includes removing duplicate values, missing values ​​and outliers; and Z-score standardization of the cleaned data.

[0089] Duplicate values ​​refer to identical data records in a dataset, which may lead to statistical bias and therefore need to be deleted. Missing values ​​refer to cases where some data fields are empty. These can be filled using imputation methods (such as using the mean, median, or prediction) to maintain data integrity. As for outliers, i.e., data points that deviate significantly from other observations, it is necessary to decide whether to remove or correct them based on the specific circumstances.

[0090] After data cleaning, the next step is to standardize the cleaned data to eliminate the influence of differences in units and numerical ranges between different features. This embodiment uses the Z-score standardization method. This standardization process ensures that different features are comparable in subsequent analysis and modeling, thereby improving algorithm performance and prediction accuracy.

[0091] In this embodiment S2, key features are extracted from the preprocessed relevant data, and a user-operation and maintenance task matrix is ​​constructed.

[0092] Specifically, key features include:

[0093] The basic information includes the tunnel structure type, length, and construction year; the geographical location information includes longitude, latitude, altitude, and climate zone information; and the historical maintenance information includes the fault type, fault frequency, maintenance measures, and maintenance costs.

[0094] Furthermore, a user-operation task matrix is ​​constructed, including:

[0095] S21. Define a user set U = {U1, U2, ..., U...}n} and the set of operation and maintenance tasks T = {T1, T2, ..., T m}; where each user U i This represents a tunnel management or maintenance team, with each maintenance task (T)... j This indicates a specific operational or maintenance requirement or project.

[0096] S22. Based on the user set U and the operation and maintenance task set T, construct the user-operation and maintenance task matrix R; where the elements R in the user-operation and maintenance task matrix R are... ij Indicates user U i For maintenance task T j Participation level;

[0097] S23. Combine the time decay function δ(t) and the weight coefficient ω to adjust the user-maintenance task matrix R, and obtain the adjusted user-maintenance task matrix R′.

[0098] The element R′ in the user-maintenance task matrix R′ ij Represented as:

[0099] R′ ij =δ(t)*ω ij *R ij

[0100] Wherein, the decay function δ(t) represents the user behavior weight at different times, δ(t) = e -λt λ represents the decay rate; weighting coefficient ω ij Indicates user U i During the execution of maintenance task T j Weighting different types of user behavior.

[0101] The time decay function δ(t) adjusts the weight of user behavior at different time points through the parameter λ (representing the decay rate), making recent behavior have a greater impact on operation and maintenance decisions, while the influence of earlier behavior gradually weakens. For example, when considering factors such as tunnel structure type, length, and construction year, recent fault types, frequencies, maintenance measures, and their costs are given higher weights to reflect the importance of these characteristics to operation and maintenance tasks under current conditions.

[0102] The weighting coefficients are used to quantify the importance of various user behaviors when performing different operation and maintenance tasks. Specifically, when handling the maintenance of a particular tunnel, the impact of fault types and frequencies may vary depending on the geographical location (such as longitude, latitude, altitude, and climate zone). The weighting coefficients help the system distinguish which behaviors (such as frequently occurring faults) are more critical to specific operation and maintenance tasks, thus giving appropriate weight when calculating user participation in operation and maintenance tasks. In this way, each element in the matrix can more accurately reflect the true relationship between users and operation and maintenance tasks, thereby improving the accuracy and applicability of operation and maintenance decisions generated based on collaborative filtering algorithms.

[0103] This step involves defining a user set and an operations and maintenance (O&M) task set, and constructing an initial user-O&M task matrix. Then, the matrix is ​​adjusted by introducing a time decay function and weighting coefficients to reflect the varying importance of user behavior across different time periods and the weight differences between different user behavior types. The final result is an adjusted user-O&M task matrix where each element comprehensively considers both time decay and behavior weighting, thus more accurately reflecting user participation in O&M tasks.

[0104] Furthermore, the user-operation task matrix is ​​stored using a sparse matrix storage format and incorporates a dynamic update mechanism; the dynamic update mechanism includes:

[0105] Acquire real-time data streams or periodically query new user behavior data;

[0106] Based on the user ID and operation and maintenance task ID in the new user behavior data, determine the affected elements in the user-operation and maintenance task matrix;

[0107] The affected elements are updated using an update function to obtain the updated user-operation task matrix.

[0108] By using a sparse matrix storage format to efficiently store the user-operation task matrix and equipped with a dynamic update mechanism, it can quickly identify the elements in the matrix that need to be updated when receiving real-time data streams or periodically queried new user behavior data, and adjust these elements through update functions, thereby reflecting the latest user behavior information in real time and ensuring the accuracy and timeliness of the operation and maintenance decision recommendation system.

[0109] In this embodiment S3, based on the user-maintenance task matrix, the similarity between users and between railway tunnel maintenance tasks is calculated using the Pearson correlation coefficient.

[0110] Calculating similarity between users includes:

[0111] S311. Define a user vector for each user; the vector includes the user's participation in all operation and maintenance tasks;

[0112] S312. Calculate the covariance and standard deviation based on the user vectors of user A and user B;

[0113] The covariance Cov(A,B) is expressed as:

[0114]

[0115] The standard deviation includes σA and σB;

[0116]

[0117] Among them, A o B o These represent the degree of participation of user A and user B in the operation and maintenance task o, respectively. represents the average engagement level of user A and user B respectively, and m represents the number of maintenance tasks shared by user A and user B;

[0118] S313. Based on the covariance and standard deviation, determine the similarity R between users using the Pearson correlation coefficient. AB :

[0119]

[0120] Among them, R AB ∈[-1,1], where 1 represents a perfect positive correlation, -1 represents a perfect negative correlation, and 0 represents no linear correlation.

[0121] Furthermore, the similarity between railway tunnel operation and maintenance tasks is calculated, including:

[0122] S321. Define an operation and maintenance task vector for each railway tunnel operation and maintenance task; the vector includes the participation degree of all users in the operation and maintenance task;

[0123] S322. Based on the operation and maintenance task vectors of operation and maintenance task A′ and operation and maintenance task B′, calculate the covariance and standard deviation.

[0124] The covariance Cov(A) ' B′) is represented as:

[0125]

[0126] The standard deviation includes σA ' and σB′;

[0127]

[0128] Among them, A′p B′ p Let A and B represent the degree of participation of user p in operation and maintenance task A' and operation and maintenance task B', respectively. These represent the user's response to maintenance task A. ' Operation and maintenance task B ' The average engagement rate, where l represents the total number of users shared by operation and maintenance tasks A′ and B′;

[0129] S323. Based on the covariance and standard deviation, determine the similarity R between users using the Pearson correlation coefficient. A′B′ :

[0130]

[0131] Among them, R A′B′ ∈[-1,1], where 1 represents a perfect positive correlation, -1 represents a perfect negative correlation, and 0 represents no linear correlation.

[0132] Furthermore, in addition to calculating the similarity between users and between railway tunnel maintenance tasks using the Pearson correlation coefficient as described above in this embodiment, other similarity calculation methods can also be used to better handle these data characteristics.

[0133] Specifically, Manhattan distance is suitable for data with many discrete values ​​or uneven data distribution, such as the construction date and altitude of railway tunnels. It is calculated by summing the absolute differences between two vectors across all dimensions to measure the distance between them. Manhattan distance is highly sensitive to discrete values ​​and unevenly distributed data, accurately reflecting the differences between these characteristics, and is suitable for multi-dimensional data analysis in railway tunnel operation and maintenance decisions.

[0134] Furthermore, in railway tunnel operation and maintenance decision recommendation systems, user participation records for different operation and maintenance tasks often exhibit high-dimensionality and sparseness. Using cosine similarity can more accurately measure the similarity between users, thereby improving the performance of the recommendation system and the user experience. In addition, cosine similarity can be efficiently computed on large-scale datasets, making it suitable for real-time recommendation systems and further enhancing the system's response speed and accuracy.

[0135] In this embodiment S4, based on the similarity between users and between railway tunnel maintenance tasks, a collaborative filtering algorithm is used to generate maintenance decisions corresponding to maintenance requests.

[0136] The collaborative filtering algorithm is used to generate operation and maintenance decisions corresponding to maintenance requests, including:

[0137] Determining operational decisions based on user-side collaborative filtering:

[0138] S411. For each user u who needs to generate operation and maintenance decisions, determine the k users most similar to user u.

[0139] N u =argmax v≠u sim(u,v)

[0140] S412. Based on the historical participation of k users in the operation and maintenance task, generate a rating for user u on the operation and maintenance task w.

[0141]

[0142] Where sim(u,v) represents the similarity between user u and user v, r v,w This indicates the degree of participation of user v in the operation and maintenance task w;

[0143] S413. Determine the operation and maintenance decision based on the score.

[0144] Furthermore, the use of collaborative filtering algorithms to generate operational decisions corresponding to maintenance requests also includes:

[0145] S421. Determine operation and maintenance decisions based on collaborative filtering on the operation and maintenance task side:

[0146] For each operation and maintenance task u′ to be decided, identify k′ operation and maintenance tasks that are most similar to operation and maintenance task u′;

[0147] N u′ =argmax v′≠u′ sim(u′,v′)

[0148] S422. Based on the user's historical participation in k′ operation and maintenance tasks, generate a rating for user w′ on operation and maintenance task u′;

[0149]

[0150] Where sim(u′,v′) represents the similarity between operation and maintenance task u′ and operation and maintenance task v′, r v′,w′ This represents the degree of participation of user w′ in operation and maintenance task v′;

[0151] S423. Determine the operation and maintenance decision based on the score.

[0152] Furthermore, to ensure that railway tunnel operation and maintenance decision recommendations meet user needs and are of high quality, a multi-stage decision-making process is adopted: First, user-side collaborative filtering is used to identify the user group most similar to the target user, and a preliminary recommendation list is generated based on these users' participation in operation and maintenance tasks; then, operation and maintenance task-side collaborative filtering is used to refine the preliminary recommendation of operation and maintenance tasks, and the best-performing tasks are selected by comparing the two scoring results; finally, these tasks are presented to the user as the final recommendation list.

[0153] This multi-stage decision-making process not only considers users' historical behavioral preferences but also incorporates the characteristics of the operational tasks themselves, ensuring the relevance and effectiveness of the recommendations. By combining information from both the user and operational task sides, recommendation bias caused by a single perspective can be effectively avoided, thus improving the overall performance of the recommendation system.

[0154] This embodiment provides a railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm. By acquiring and preprocessing basic information, geographical location information, and historical maintenance information of the railway tunnel, key features are extracted and a user-maintenance task matrix is ​​constructed. The similarity between users and maintenance tasks is calculated using the Pearson correlation coefficient, and then a collaborative filtering algorithm is used to generate operation and maintenance decisions corresponding to maintenance requests. This method not only improves the efficiency and accuracy of railway tunnel operation and maintenance work but also promotes a proactive shift in maintenance work, facilitating preventative maintenance and optimizing resource allocation, ultimately improving the operational safety and reliability of railway tunnels. By combining information from both the user and maintenance task sides through a multi-stage decision-making process, the recommended results are ensured to meet user needs while maintaining high quality.

[0155] Example 2;

[0156] like Figure 2 As shown, this embodiment provides a railway tunnel operation and maintenance decision recommendation system based on collaborative filtering algorithm, including:

[0157] Data acquisition and preprocessing module: used to acquire relevant data from multiple railway tunnels and perform preprocessing; the relevant data includes: basic information, geographical location information, and historical maintenance information;

[0158] Feature extraction and matrix construction module: used to extract key features from preprocessed relevant data and construct a user-operation task matrix;

[0159] Similarity calculation module: used to calculate the similarity between users and between railway tunnel maintenance tasks based on the user-maintenance task matrix and the Pearson correlation coefficient;

[0160] Decision generation module: Based on the similarity between users and between railway tunnel maintenance tasks, it uses a collaborative filtering algorithm to generate maintenance decisions corresponding to maintenance requests.

[0161] This embodiment provides a railway tunnel operation and maintenance decision recommendation system based on collaborative filtering algorithm, including a data acquisition and preprocessing module, a feature extraction and matrix construction module, a similarity calculation module, and a decision generation module. By acquiring and preprocessing basic information, geographical location information, and historical maintenance information of railway tunnels, key features are extracted and a user-operation and maintenance task matrix is ​​constructed. The similarity between users and operation and maintenance tasks is calculated using the Pearson correlation coefficient. Finally, the collaborative filtering algorithm is used to generate operation and maintenance decisions corresponding to maintenance requests, thereby improving the efficiency and accuracy of railway tunnel operation and maintenance work and promoting the initiative and prevention of maintenance work.

[0162] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0163] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm, characterized in that, Includes the following steps: S1. Acquire relevant data from multiple railway tunnels and perform preprocessing; the relevant data includes: basic information, geographical location information, and historical maintenance information; S2. Extract key features from the preprocessed data and construct a user-operation task matrix; S3. Based on the user-maintenance task matrix, calculate the similarity between users and between railway tunnel maintenance tasks using the Pearson correlation coefficient; S4. Based on the similarity between users and between railway tunnel maintenance tasks, a collaborative filtering algorithm is used to generate maintenance decisions corresponding to maintenance requests.

2. The railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 1, wherein the key features in S2 include: The basic information includes the tunnel structure type, length, and construction year; Geographic location information includes longitude, latitude, altitude, and climate zone information; The fault type, fault frequency, maintenance measures, and maintenance costs are included in the historical maintenance information.

3. The railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 1, characterized in that, In S2, the user-operation and maintenance task matrix is ​​constructed, including: S21. Define a user set U = {U1, U2, ..., U...} n } and the set of operation and maintenance tasks T = {T1, T2, ..., T m }; where each user U i This represents a tunnel management or maintenance team, with each maintenance task (T)... j This indicates a specific operational or maintenance requirement or project. S22. Based on the user set U and the operation and maintenance task set T, construct the user-operation and maintenance task matrix R; where the elements R in the user-operation and maintenance task matrix R are... ij Indicates user U i For maintenance task T j Participation level; S23. Combine the time decay function δ(t) and the weight coefficient ω to adjust the user-maintenance task matrix R, and obtain the adjusted user-maintenance task matrix R′.

4. The railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 3, characterized in that, In S23, the element R′ in the user-maintenance task matrix R′ ij Represented as: R′ ij =δ(t)*ω ij *R ij Wherein, the decay function δ(t) represents the user behavior weight at different times, δ(t) = e -λt λ represents the decay rate; weighting coefficient ω ij Indicates user U i During the execution of maintenance task T j Weighting different types of user behavior.

5. The railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 1, characterized in that, In S2, the user-operation and maintenance task matrix is ​​stored in a sparse matrix storage format and combined with a dynamic update mechanism. The dynamic update mechanism includes: Acquire real-time data streams or periodically query new user behavior data; Based on the user ID and operation and maintenance task ID in the new user behavior data, determine the affected elements in the user-operation and maintenance task matrix; The affected elements are updated using an update function to obtain the updated user-operation task matrix.

6. The railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 1, characterized in that, In step S3, calculating the similarity between users includes: S311. Define a user vector for each user; the vector includes the user's participation in all operation and maintenance tasks; S312. Calculate the covariance and standard deviation based on the user vectors of user A and user B; The covariance Cov(A,B) is expressed as: The standard deviation includes σA and σB; Among them, A o B o These represent the degree of participation of user A and user B in the operation and maintenance task o, respectively. represents the average engagement level of user A and user B respectively, and m represents the number of maintenance tasks shared by user A and user B; S313. Based on the covariance and standard deviation, determine the similarity R between users using the Pearson correlation coefficient. AB : Among them, R AB ∈[-1,1], where 1 represents a perfect positive correlation, -1 represents a perfect negative correlation, and 0 represents no linear correlation.

7. A railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 7, characterized in that, In step S3, the similarity between railway tunnel maintenance tasks is calculated, including: S321. Define an operation and maintenance task vector for each railway tunnel operation and maintenance task; the vector includes the participation degree of all users in the operation and maintenance task; S322. Based on the operation and maintenance task vectors of operation and maintenance task A′ and operation and maintenance task B′, calculate the covariance and standard deviation. The covariance Cov(A′,B′) is expressed as: The standard deviations include σA' and σB′; Among them, A′ p B′ p These represent the degree of participation of user p in operation and maintenance task A′ and operation and maintenance task B′, respectively. represents the average user participation in operation and maintenance task A′ and operation and maintenance task B′, respectively, and l represents the total number of users in operation and maintenance task A′ and operation and maintenance task B′. S323. Based on the covariance and standard deviation, determine the similarity R between users using the Pearson correlation coefficient. A′B′ : Among them, R A′B' ∈[-1,1], where 1 represents a perfect positive correlation, -1 represents a perfect negative correlation, and 0 represents no linear correlation.

8. A railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 7, characterized in that, In step S4, a collaborative filtering algorithm is used to generate operation and maintenance decisions corresponding to the maintenance request, including: Determining operational decisions based on user-side collaborative filtering: S411. For each user u who needs to generate operation and maintenance decisions, determine the k users most similar to user u. N u =argmax v≠u sim(u,v) S412. Based on the historical participation of k users in the operation and maintenance task, generate a rating for user u on the operation and maintenance task w. Where sim(u,v) represents the similarity between user u and user v, r v,w This indicates the degree of participation of user v in the operation and maintenance task w; S413. Determine the operation and maintenance decision based on the score.

9. A railway tunnel operation and maintenance decision recommendation method based on collaborative filtering algorithm according to claim 7, characterized in that, In step S4, the collaborative filtering algorithm is used to generate operation and maintenance decisions corresponding to maintenance requests, which also includes: S421. Determine operation and maintenance decisions based on collaborative filtering on the operation and maintenance task side: For each operation and maintenance task u′ to be decided, identify k′ operation and maintenance tasks that are most similar to operation and maintenance task u′. N u′ =argmax v'≠u′ sim(u′,v′) S422. Based on the user's historical participation in k′ operation and maintenance tasks, generate a rating for user w′ on operation and maintenance task u′; Where sim(u',v′) represents the similarity between operation and maintenance task u′ and operation and maintenance task v′, r v′,w′ This represents the degree of participation of user w′ in operation and maintenance task v′; S423. Determine the operation and maintenance decision based on the score.

10. A railway tunnel operation and maintenance decision recommendation system based on collaborative filtering algorithm, characterized in that, include: Data acquisition and preprocessing module: used to acquire relevant data from multiple railway tunnels and perform preprocessing; The relevant data includes: basic information, geographic location information, and historical maintenance information; Feature extraction and matrix construction module: used to extract key features from preprocessed relevant data and construct a user-operation task matrix; Similarity calculation module: used to calculate the similarity between users and between railway tunnel maintenance tasks based on the user-maintenance task matrix and the Pearson correlation coefficient; Decision generation module: Based on the similarity between users and between railway tunnel maintenance tasks, it uses a collaborative filtering algorithm to generate maintenance decisions corresponding to maintenance requests.

Images