A method for recommending oilfield cloud application components based on coefficient of variation method

Abstract

The present invention discloses a method for recommending oilfield cloud application components based on a coefficient of variation method, comprising the following steps: S1, collecting oilfield cloud application component log data; S2, data normalization processing; normalizing an oilfield application component evaluation matrix to eliminate dimensional differences between different indicators, so that the numerical value of each indicator is mapped to an interval between [0, 1], and obtaining a standardized evaluation matrix; S3, constructing a virtual rating matrix by using the coefficient of variation method; S4, establishing a recommendation model; establishing an oilfield cloud application component recommendation model by combining a user-based collaborative filtering algorithm with a coefficient of variation method, adopting the coefficient of variation method to analyze and simulate user ratings from multiple dimensions, calculating the similarity between oilfield cloud users according to the Jaccard formula, calculating the user's interest in the oilfield cloud application component by using an interest calculation formula, and the interest value is the user's recommendation result.

Description

Technical Field

[0001] The present invention relates to the field of oilfield cloud technology, and in particular to a method for recommending oilfield cloud application components based on a coefficient of variation method. Background Art

[0002] In recent years, as oilfield enterprise information technology has become increasingly sophisticated, the volume of application component log data generated daily by oilfield cloud computing has become enormous, leading to an increasingly serious information overload problem. How users can quickly find the application components they need within this vast pool of oilfield cloud application components is an urgent issue facing oilfield cloud computing.

[0003] As we all know, countless scientists and engineers have proposed numerous promising solutions to address information overload. With the advent of the big data era, oilfield cloud data is growing exponentially. Faced with unprecedented data volumes, traditional recommendation technologies are evolving towards integration with big data. Currently, various new recommendation technologies remain a hot topic in academic research.

[0004] Traditional recommendation methods build recommendation models based on user rating data and calculate user demand for items. However, since Oilfield Cloud application components lack explicit user feedback data, meaning user ratings of application components are missing, it's impossible to directly build recommendation models based on user ratings. Simulating user ratings on a 0-1 scale based on Oilfield Cloud user usage records ignores user usage frequency and the application scope of the application component, resulting in a certain degree of bias in the recommendation results and inaccurate results. Summary of the Invention

[0005] The purpose of the present invention is to address the problems faced by traditional recommendation methods in oilfield cloud application components. The present invention provides a new oilfield cloud application component recommendation method based on the coefficient of variation method.

[0006] In order to achieve the above object, the technical solution of the present invention is:

[0007] A method for recommending oilfield cloud application components based on a coefficient of variation method includes the following steps:

[0008] S1. Oilfield cloud application component log data collection: Use the oilfield application component log to collect user application component usage records. Through filtering, obtain the user's access volume to each application component and the application scope of each component itself. Use these two indicators to construct the oilfield application component evaluation matrix X for each user:

[0009]

[0010] Where m represents m indicators, and n represents n oilfield application components;

[0011] S2. Data normalization: normalizing the oilfield application component evaluation matrix to eliminate the dimensional differences between different indicators, mapping the values ​​of each indicator to the interval [0, 1], and obtaining a normalized evaluation matrix;

[0012] S3. Use the coefficient of variation method to construct a virtual scoring matrix. The coefficient of variation method is used to measure the degree of difference in the values ​​of each indicator. The information contained in each indicator is directly used to calculate the weight of the indicator. First, the average value and standard deviation of each evaluation indicator are calculated, then the coefficient of variation of each indicator is calculated, and then the weight of each indicator is calculated. Finally, the total score of each oilfield application component is calculated. The calculation formula is as follows:

[0013]

[0014]

[0015]

[0016]

[0017]

[0018] Where V i is the coefficient of variation of the i-th indicator, σ i is the standard deviation of the i-th indicator, is the average value of the i-th indicator, w i is the weight of each indicator, r j The total score for each oilfield application component is also used as the user's simulated score for each oilfield application component;

[0019] S4. Establishment of a recommendation model: A user-based collaborative filtering algorithm is combined with a coefficient of variation method to establish an oilfield cloud application component recommendation model. The coefficient of variation method is used to analyze and simulate user ratings from multiple dimensions. The similarity between oilfield cloud users is calculated according to the Jaccard formula. The user's interest in the oilfield cloud application component is calculated using the interest calculation formula. The interest value is the user's recommendation result. The Jaccard similarity calculation formula is as follows:

[0020]

[0021] The formula for calculating interest is as follows:

[0022]

[0023] Where N(u) represents the set of application components used by user u, N(v) represents the set of application components used by user v, and w uvrepresents the similarity between user u and user v, N(i) represents the set of users who have used application component i, r vi It represents the virtual rating of user v to application component i; the interest value is the user's recommendation result.

[0024] Furthermore, the normalization method in step S2 is a deviation normalization method, and the conversion formula is as follows:

[0025]

[0026] Among them, max is the maximum value of each indicator data, and min is the minimum value of each indicator data.

[0027] Compared with the prior art, the present invention has the following advantages and positive effects:

[0028] The present invention is based on oilfield cloud log data and adopts the coefficient of variation method to evaluate application components with usage frequency and application scope as evaluation indicators, thereby simulating user ratings based on the evaluation results, and combining the traditional user-based collaborative filtering algorithm to calculate the user's demand for oilfield cloud application components. It can effectively solve the problem of deviation in recommendation methods caused by the lack of explicit feedback data, greatly improve the accuracy of recommendations, and bring great help to users in selecting oilfield application components. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0030] Figure 1 It is a process framework diagram of the present invention. DETAILED DESCRIPTION

[0031] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts, any modifications, equivalent replacements, improvements, etc., shall be included in the scope of protection of the present invention.

[0032] like Figure 1 As shown in Figure 2, the oilfield cloud application component recommendation method based on the coefficient of variation method mainly includes the following steps:

[0033] 1. Oilfield cloud application component log data collection

[0034] We used Python's regularization module to preprocess the oilfield cloud log data, extracting attribute data such as user name, application component, and user unit. We then organized the data by user name and stored it in a two-dimensional list. We filtered the data to obtain the number of visits to each application component and the application scope of each component. Using these two metrics, we constructed an oilfield application component evaluation matrix X for each user:

[0035]

[0036] Where m represents m indicators, and n represents n oilfield application components;

[0037] 2. Data normalization

[0038] The oilfield application component evaluation matrix is ​​normalized to eliminate the dimensional differences between different indicators. The normalization method can use the deviation standardization method to map the values ​​of each indicator to the interval [0, 1] to obtain a normalized evaluation matrix. The conversion formula is as follows:

[0039]

[0040] Among them, max is the maximum value of each indicator data, and min is the minimum value of each indicator data;

[0041] 3. Calculate the virtual scoring matrix using the coefficient of variation method

[0042] The coefficient of variation method is used to measure the degree of difference in the values ​​of each indicator. The information contained in each indicator is directly used to calculate the weight of the indicator. First, the average value and standard deviation of each evaluation indicator are calculated, then the coefficient of variation of each indicator is calculated, and then the weight of each indicator is calculated. Finally, the total score of each oilfield application component is calculated, thereby obtaining a virtual scoring matrix for each component by each user. The calculation formula is as follows:

[0043]

[0044]

[0045]

[0046]

[0047]

[0048] Where Vi is the coefficient of variation of the i-th indicator, σi is the standard deviation of the i-th indicator, is the average value of the i-th indicator, wi is the weight of each indicator, rj is the total score of each oilfield application component, and is also the user's simulated score for each oilfield application component;

[0049] 4. Establishment of recommendation model

[0050] The oilfield cloud application component recommendation model is established by combining the traditional user-based collaborative filtering algorithm with the coefficient of variation method. The oilfield cloud log lacks user rating data on application components. If the traditional 0, 1 simulation user rating is used based on the user's usage record, the user's usage frequency and the application scope of the application component are ignored. The coefficient of variation method is used to simulate user ratings, and the virtual rating matrix is ​​used as the input of the Jaccard formula. The similarity between oilfield cloud users is calculated according to the Jaccard formula, and the similarity is input into the user-based collaborative filtering recommendation algorithm. The demand calculation formula is used to calculate the user's demand for oilfield cloud application components. The user's demand for oilfield cloud application components is calculated one by one. The Jaccard similarity calculation formula is as follows:

[0051]

[0052] The formula for calculating demand is as follows:

[0053]

[0054] Where N(u) represents the set of application components used by user u, N(v) represents the set of components used by user v, wuv represents the similarity between user u and user v, N(i) represents the set of users who have used application component i, and rvi represents the virtual rating of application component i by user v.

[0055] 5. Actual testing

[0056] Obtain oilfield cloud application data and input it into the oilfield cloud application recommendation method based on the coefficient of variation method. The output is the user's recommendation result.

[0057] The present invention is based on oilfield cloud log data and adopts the coefficient of variation method to evaluate application components with usage frequency and application scope as evaluation indicators, thereby simulating user ratings based on the evaluation results, and combining the traditional user-based collaborative filtering algorithm to calculate the user's demand for oilfield cloud application components. It can effectively solve the problem of deviation in recommendation methods caused by the lack of explicit feedback data, greatly improve the accuracy of recommendations, and bring great help to users in selecting oilfield application components.

Claims

1. A method for recommending oilfield cloud application components based on the coefficient of variation method, characterized by: The following steps are involved: S1. Oilfield cloud application component log data collection; The oilfield application component logs are used to collect user application component usage records. Through filtering, the user's access volume to each application component and the application scope of each component are obtained. These two indicators are used to construct the oilfield application component evaluation matrix X for each user: Where m represents m indicators, and n represents n oilfield application components; S2. Data normalization: normalize the oilfield application component evaluation matrix to eliminate the dimensional differences between different indicators, map the values ​​of each indicator to the interval [0, 1], and obtain a standardized evaluation matrix; S3. Use the coefficient of variation method to construct a virtual scoring matrix. The coefficient of variation method is used to measure the degree of difference in the values ​​of each indicator. The information contained in each indicator is directly used to calculate the weight of the indicator. First, the average value and standard deviation of each evaluation indicator are calculated, then the coefficient of variation of each indicator is calculated, and then the weight of each indicator is calculated. Finally, the total score of each oilfield application component is calculated. The calculation formula is as follows: Where V i is the coefficient of variation of the i-th indicator, σ i is the standard deviation of the i-th indicator, is the average value of the i-th indicator, w i is the weight of each indicator, r j The total score for each oilfield application component is also used as the user's simulated score for each oilfield application component; S4. Establishment of a recommendation model: A user-based collaborative filtering algorithm is combined with a coefficient of variation method to establish an oilfield cloud application component recommendation model. The coefficient of variation method is used to analyze and simulate user ratings from multiple dimensions. The similarity between oilfield cloud users is calculated according to the Jaccard formula. The user's interest in the oilfield cloud application component is calculated using the interest calculation formula. The interest value is the user's recommendation result. The Jaccard similarity calculation formula is as follows: The formula for calculating interest is as follows: Where N(u) represents the set of application components used by user u, N(v) represents the set of application components used by user v, and w uv represents the similarity between user u and user v, N(i) represents the set of users who have used application component i, r vi represents the virtual rating of user v on application component i; The normalization method in step S2 is a deviation normalization method, and the conversion formula is as follows: Among them, max is the maximum value of each indicator data, and min is the minimum value of each indicator data.

Images