News content propagation prediction method and system based on big data

Abstract

The invention relates to the technical field of news transmission prediction, and particularly discloses a news content transmission prediction method and system based on big data, and the method comprises the steps: determining the information transfer degree of a user according to the activeness of a group where the user joins and the number of people in the group; determining a selection probability of each user according to the information transfer degree, randomly selecting the users based on the selection probability until the number of the users reaches a preset threshold value, and taking the selected users as reference users; acquiring browsing content of a reference user, identifying the browsing content, and creating a keyword group; when news content containing a prediction request is received, predicting the spreading degree of the news content based on the keyword group; the propagation degree is used for representing the propagation capability of the news content, and the higher the propagation degree is, the stronger the propagation capability is; the method is essentially a propagation prediction scheme based on the exposure rate, the scheme does not involve a complex algorithm, and a prediction result with high accuracy can be obtained with low cost.

Description

Technical Field

[0001] The present invention relates to the technical field of news dissemination prediction, and in particular to a news content dissemination prediction method and system based on big data. Background Art

[0002] In an environment where smart devices are extremely popular, everyone will upload some information and share their emotions and other status with others. In addition, with the advent of the self-media era, some high-quality information can often obtain a certain amount of traffic and then be monetized.

[0003] Under the existing technical background, it is difficult for the information uploader to know the dissemination capacity of the information he uploaded. Whether he can obtain higher traffic depends entirely on his own feeling, which is highly accidental. How to predict the dissemination process of information while uploading information, and then let users modify their own tags to obtain higher traffic is the technical problem that the technical solution of the present invention aims to solve. Summary of the invention

[0004] The purpose of the present invention is to provide a news content dissemination prediction method and system based on big data to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method and system for predicting news content dissemination based on big data, the method comprising:

[0007] Receive information access permissions granted by the user, query the contacts of each user based on the information access permissions, and simultaneously query the activity and number of groups that the user has joined;

[0008] Determine the user's information transfer degree based on the activity of the group the user joins and the number of group members;

[0009] Determine the selection probability of each user according to the information transfer degree, randomly select users based on the selection probability until the number of users reaches a preset threshold, and use the selected users as reference users;

[0010] Obtain browsing content of a reference user, identify the browsing content, and create a keyword group; the keyword group contains keyword items and weight items;

[0011] When news content containing a prediction request is received, the spread degree of the news content is predicted based on the keyword group; the spread degree is used to characterize the spread ability of the news content, and the higher the spread degree, the stronger the spread ability.

[0012] As a further solution of the present invention, the steps of receiving the information acquisition authority granted by the user, querying the contacts of each user based on the information acquisition authority, and simultaneously querying the activity and number of groups joined by the user include:

[0013] Receive information access permissions granted by the user, and query the contacts of each user based on the information access permissions;

[0014] Locate the group in the contacts, obtain the number of members in the group and the total number of messages in each time period;

[0015] Calculate the activity of the group based on the number of people in the group and the total number of messages in each time period;

[0016] The activity level is:

[0017] In the formula, H represents the activity, N represents the number of people in the group, N0 is the preset number of people, α, β and γ are the preset correction coefficients, and X max is the maximum number of messages in each period, X e is the average number of information in each period, T σ is the standard deviation of the total number of messages in each period.

[0018] As a further solution of the present invention: the step of determining the user's information transfer degree according to the activity of the group the user joins and the number of people in the group includes:

[0019] Obtain the time when a user sends information in each group based on the information acquisition authority;

[0020] For any group, the affinity between the user and the group is calculated based on the time when the message was sent;

[0021] The information transfer degree is calculated based on the user's affinity with each group and the activity of the corresponding group;

[0022] The affinity calculation process is as follows:

[0023]

[0024] The calculation process of the information transfer degree is:

[0025]

[0026] In the formula, Q represents affinity, N is the total number of information sending times, and t i represents the time when the i-th information is sent, X(t i ) represents the total number of messages in the time period corresponding to the i-th message sending time, X(t i-1) represents the total number of messages in the time period corresponding to the previous message sending time of the i-th message sending time, X(t i+1 ) represents the total number of messages in the time period corresponding to the next message sending time after the i-th message sending time; X(t0) and X(t N+1 ) is the default value; Z represents the information transfer degree, M represents the total number of user groups, and H j represents the activity of the jth group, Q j represents the affinity between the user and the jth group; C1, C2 and C3 are all preset constants; X(0) represents the preset number threshold.

[0027] As a further solution of the present invention: the step of determining the selection probability of each user according to the information transfer degree, randomly selecting users based on the selection probability until the number of users reaches a preset threshold, and using the selected users as reference users includes:

[0028] Calculating the selection probability corresponding to the information transfer degree according to the direct proportion of the information transfer degree;

[0029] Randomly select information transfer degrees based on the selection probability, cyclically select and record the number of information transfer degrees, and exit the loop when the number of information transfer degrees reaches a preset threshold;

[0030] A range is determined based on the information transfer degree and a preset ratio, and users in the range are queried, and a user is randomly selected as a reference user.

[0031] As a further solution of the present invention: the step of obtaining the browsing content of the reference user, identifying the browsing content, and creating a keyword group includes:

[0032] Query the browsing content of reference users within a preset time based on information acquisition permissions;

[0033] When the browsing content is audio information, performing audio recognition on it and converting it into text data;

[0034] When the browsing content is image information, the image information is recognized based on AI and converted into text data;

[0035] Extract keywords from text data to obtain keyword groups;

[0036] Compare and merge the keyword groups corresponding to all browsed contents, calculate the weight according to the number of groups corresponding to each keyword, and generate the weight item.

[0037] As a further solution of the present invention: when receiving news content containing a prediction request, the step of predicting the spread of the news content based on the keyword group includes:

[0038] When receiving news content containing a prediction request, extracting keywords of the news content based on the same recognition scheme as for browsing content;

[0039] Compare the extracted keywords with keyword groups to determine the degree of spread;

[0040] The calculation process of propagation degree is:

[0041] In the formula, S represents the propagation degree, B k is the weight of the keyword corresponding to the kth keyword in the keyword group; C4 is the preset correction coefficient.

[0042] The technical solution of the present invention also provides a news content dissemination prediction system based on big data, the system comprising:

[0043] The information query module is used to receive the information acquisition authority granted by the user, query the contacts of each user based on the information acquisition authority, and simultaneously query the activity and number of groups joined by the user;

[0044] The transfer degree calculation module is used to determine the user's information transfer degree according to the activity of the group the user joins and the number of people in the group;

[0045] A user selection module, used to determine the selection probability of each user according to the information transfer degree, randomly select users based on the selection probability until the number of users reaches a preset threshold, and use the selected users as reference users;

[0046] A browsing content identification module is used to obtain the browsing content of the reference user, identify the browsing content, and create a keyword group; the keyword group contains keyword items and weight items;

[0047] The spread degree prediction module is used to predict the spread degree of the news content based on the keyword group when receiving news content containing a prediction request; the spread degree is used to characterize the spread ability of the news content, and the higher the spread degree, the stronger the spread ability.

[0048] As a further solution of the present invention: the information query module includes:

[0049] A contact query unit, used to receive information acquisition permissions granted by a user, and query contacts of each user based on the information acquisition permissions;

[0050] A group information acquisition unit locates the group in the contacts, obtains the number of group members and the total number of information in each time period;

[0051] An activity calculation unit, used to calculate the activity of the group according to the number of members in the group and the total number of messages in each time period;

[0052] The activity level is:

[0053] In the formula, H represents the activity, N represents the number of people in the group, N0 is the preset number of people, α, β and γ are the preset correction coefficients, and X max is the maximum number of messages in each period, X e is the average number of information in each period, T σ is the standard deviation of the total number of messages in each period.

[0054] As a further solution of the present invention: the transit degree calculation module includes:

[0055] A time query unit, used to obtain the information sending time of the user in each group based on the information acquisition authority;

[0056] An affinity calculation unit, used for calculating the affinity between the user and any group according to the time when the information is sent;

[0057] The information transfer degree is calculated based on the user's affinity with each group and the activity of the corresponding group;

[0058] The affinity calculation process is as follows:

[0059]

[0060] The calculation process of the information transfer degree is:

[0061]

[0062] In the formula, Q represents affinity, N is the total number of information sending times, and t i represents the time when the i-th information is sent, X(t i ) represents the total number of messages in the time period corresponding to the i-th message sending time, X(t i-1 ) represents the total number of messages in the time period corresponding to the previous message sending time of the i-th message sending time, X(t i+1 ) represents the total number of messages in the time period corresponding to the next message sending time after the i-th message sending time; X(t0) and X(t N+1 ) is the default value; Z represents the information transfer degree, M represents the total number of user groups, and H j represents the activity of the jth group, Q j represents the affinity between the user and the jth group; C1, C2 and C3 are all preset constants; X(0) represents the preset number threshold.

[0063] As a further solution of the present invention: the user selection module includes:

[0064] A selection probability calculation unit, used to calculate the selection probability corresponding to the information transfer degree according to the direct proportion of the information transfer degree;

[0065] A loop selection unit, used for randomly selecting information transfer degrees based on a selection probability, cyclically selecting and recording the number of information transfer degrees, and exiting the loop when the number of information transfer degrees reaches a preset threshold;

[0066] The user query unit is used to determine a range based on the information transfer degree and a preset ratio, query users in the range, and randomly select a user as a reference user.

[0067] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention analyzes each user, and can calculate the user's information dissemination ability based on the activity of all groups the user joins and the number of groups, and then screens out users with strong information dissemination ability, creates keyword groups based on the preference data of such users, and then predicts the dissemination of news content. It is essentially a dissemination prediction scheme based on exposure rate, which does not involve complex algorithms and can obtain highly accurate prediction results at a relatively low cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0069] Figure 1 The overall flow chart of the news content dissemination prediction method based on big data is shown.

[0070] Figure 2 The structural diagram of the news content dissemination prediction system based on big data is shown. DETAILED DESCRIPTION

[0071] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0072] Figure 1 The following is a general flow chart of a news content propagation prediction method and system based on big data. In an embodiment of the present invention, a news content propagation prediction method based on big data includes:

[0073] Step S100: receiving the information acquisition authority granted by the user, querying the contacts of each user based on the information acquisition authority, and simultaneously querying the activity and number of groups joined by the user;

[0074] The technical solution of the present invention involves the query process of contacts and information, which requires obtaining permissions in advance. It should be noted that this permission is often obtained when a user registers an account. In fact, the premise for providing social services is to have this permission, and then convert the touch screen information input by the user into audio-visual information and forward it to another person; for example, when using chat software, the provider of the chat software must be able to obtain all the user's information. Without this information, the service cannot be provided at all. It should be noted that this permission is limited to the service provider's information acquisition permission, not information forwarding permission.

[0075] Based on the information acquisition permissions, query each user's contacts, query the groups that the user has joined in the contacts, and obtain the group's activity and group size. The group's activity is used to characterize the frequency of information interaction in the group, and the group size is the number of people in the group.

[0076] Step S200: determining the information transfer degree of the user according to the activity of the group the user joins and the number of people in the group;

[0077] The user's information dissemination ability can be calculated based on the activity and number of group members of all the groups the user joins. This is represented by the parameter "information transfer degree". The higher the information transfer degree, the stronger the dissemination ability. The information transfer degree is generally proportional to the activity and the number of group members. In addition, the parameter of the user's sent information volume can also be introduced. The greater the amount of information sent, the greater the information transfer degree.

[0078] Step S300: determining the selection probability of each user according to the information transfer degree, randomly selecting users based on the selection probability until the number of users reaches a preset threshold, and using the selected users as reference users;

[0079] Information transfer degree is a characteristic of users. The selection probability of each user can be calculated based on the information transfer degree. The purpose of this process is to screen out users who are worthy of reference from all users. Since the number of reference users is a preset finite value, the users who are most worthy of analysis can be selected based on the selection probability, that is, the reference users.

[0080] Since the selection process is based on the selection probability, a low selection probability does not mean that it will not be selected at all. Compared with the traditional selection method based on a fixed threshold (comparing the information transfer degree with the fixed threshold and then selecting), this method is softer and more comprehensive.

[0081] Step S400: Obtain browsing content of a reference user, identify the browsing content, and create a keyword group; the keyword group contains keyword items and weight items;

[0082] The browsing content of the reference user is obtained. This process also needs to be completed based on the information acquisition authority. Of course, the user can revoke the authority in this process, that is, the incognito browsing mode, so that the background will not record the user's browsing content; if the reference user does not grant authority, then some reference users can be selected based on step S300, or the reference users who do not grant authority can be eliminated from the reference users. Both solutions are feasible solutions; the browsing content can be identified to extract keywords, and the weight of each keyword can be calculated according to the number of occurrences of the keyword. The keywords and their weights are counted to obtain keyword groups.

[0083] Step S500: when receiving news content containing a prediction request, predicting the spread degree of the news content based on the keyword group; the spread degree is used to characterize the spread ability of the news content, and the higher the spread degree, the stronger the spread ability;

[0084] When news content containing a prediction request is received, the recognition scheme for browsing content in step S400 is used to identify the news content, extract keywords, and then place the extracted keywords in a keyword group for comparison, thereby predicting the degree of spread of the news content; the degree of spread is used to characterize the spread ability of the news content, and the higher the degree of spread, the stronger the spread ability.

[0085] The practical significance of the above scheme is to locate the group of people who have the ability to spread news on a large scale among all users, obtain the browsing content of this group of people, extract keywords from the browsing content, and use the extracted keywords as the best content in the current situation. When a prediction request is received, the news content is compared with the best content. The closer the comparison result is, the easier it is for the news content to be viewed by users with strong dissemination capabilities. Correspondingly, it is regarded as having stronger dissemination capabilities and higher dissemination degree.

[0086] It should be noted that, during the application phase of the technical solution of the present invention, the management party also needs to preset a time range, such as one week. Taking the current moment as the starting point, the activity level and number of group members joined by users in the previous week are obtained, and the degree of dissemination finally calculated corresponds to the degree of dissemination of this week. As time goes by, the activity level and number of group members joined by users in the previous week change, and the degree of dissemination also changes, thereby ensuring timeliness.

[0087] Regarding step S100, the steps of receiving the information acquisition authority granted by the user, querying the contacts of each user based on the information acquisition authority, and simultaneously querying the activity and number of groups joined by the user include:

[0088] Receive information access permissions granted by the user, and query the contacts of each user based on the information access permissions;

[0089] Locate the group in the contacts, obtain the number of members in the group and the total number of messages in each time period;

[0090] The activity of the group is calculated based on the number of members in the group and the total number of messages in each time period.

[0091] The above content describes the group analysis process in detail. The process of granting information acquisition permissions is the initial step, in which the user grants information acquisition permissions; based on the information acquisition permissions, the contacts of each user are queried, the group is located in the contacts, and the number of group members in the located group and the total number of information in each time period are queried. The time period is in minutes, such as one minute or five minutes; the activity level can be calculated based on the number of group members and the total number of information in each time period.

[0092] The activity level is:

[0093] In the formula, H represents the activity, N represents the number of people in the group, N0 is the preset number of people, α, β and γ are the preset correction coefficients, and X max is the maximum number of messages in each period, X e is the average number of information in each period, T σ is the standard deviation of the total number of messages in each period.

[0094] The calculation principle of activity is not complicated. max , X e and T σ are positively proportional, It is a correction item. If the number of people in the group is greater than the preset threshold, the value is greater than 1, otherwise it is less than 1. In order to prevent negative values ​​from occurring in the activity level, all negative values ​​can be set to zero.

[0095] Regarding step S200, the step of determining the user's information transfer degree according to the activity of the group the user joins and the number of group members includes:

[0096] Obtain the time when a user sends information in each group based on the information acquisition authority;

[0097] For any group, the affinity between the user and the group is calculated based on the time when the message was sent;

[0098] The information transfer degree is calculated based on the user's affinity with each group and the activity of the corresponding group.

[0099] Users occasionally speak in various groups. The time when users speak in various groups is obtained based on information acquisition permissions, which is called the information sending time. For any group, the activity of the group is read, and the affinity between the user and the group is calculated based on the information sending time. The information transfer degree can be calculated by combining the two parameters of affinity and activity. Among them, affinity is used to represent the preference between users and groups. The higher the affinity, the higher the preference and the more they like to speak.

[0100] The affinity calculation process is as follows:

[0101] The affinity calculation process is as follows:

[0102]

[0103] The calculation process of the information transfer degree is:

[0104]

[0105] In the formula, Q represents affinity, N is the total number of information sending times, and t i represents the time when the i-th information is sent, X(t i ) represents the total number of messages in the time period corresponding to the i-th message sending time, X(t i-1 ) represents the total number of messages in the time period corresponding to the previous message sending time of the i-th message sending time, X(t i+1 ) represents the total number of messages in the time period corresponding to the next message sending time after the i-th message sending time; X(t0) and X(t N+1 ) is the default value; Z represents the information transfer degree, M represents the total number of user groups, and H j represents the activity of the jth group, Q j represents the affinity between the user and the jth group; C1, C2 and C3 are all preset constants; X(0) represents the preset number threshold.

[0106] Affinity is related to the total number of messages in each time period. If there are many messages in the group before and after a user's speech, the speech is considered to be more important. If there are few messages before and after a user's speech, the speech is considered not to be very important. The affinity is obtained by accumulating all the calculation results reflecting whether the speech is important.

[0107] After the user's affinity with each group is calculated, the product of the affinity and the activity is calculated and then accumulated. The resulting value is called the information transfer degree.

[0108] Regarding step S300, the step of determining the selection probability of each user according to the information transfer degree, randomly selecting users based on the selection probability until the number of users reaches a preset threshold, and using the selected users as reference users includes:

[0109] Calculating the selection probability corresponding to the information transfer degree according to the direct proportion of the information transfer degree;

[0110] Randomly select information transfer degrees based on the selection probability, cyclically select and record the number of information transfer degrees, and exit the loop when the number of information transfer degrees reaches a preset threshold;

[0111] A range is determined based on the information transfer degree and a preset ratio, and users in the range are queried, and a user is randomly selected as a reference user.

[0112] The selection probability is determined based on the direct proportion of the information transfer degree. After the selection probability is determined, a selection cycle is created to select the information transfer degree based on the selection probability until the number of cycles reaches a preset number. At this time, multiple selected information transfer degrees are obtained. This process seems redundant because when the number of selections is infinite, the selection results of the information transfer degree will tend more and more towards the selection probability distribution. However, the number of cycles in the application scenario of the present application is limited. Therefore, the selected information transfer degree has a higher randomness. In actual situations, the propagation process itself has a certain degree of randomness. This solution is more in line with the actual situation.

[0113] Finally, take the information transfer degree as the benchmark value, with a fluctuation of 5% (5% is the ratio, and the fluctuation interval is called the range), select users whose information transfer degree is within the range (information transfer degree is a label of the user), and randomly select one of the selected users as the reference user.

[0114] The above scheme is actually a gradient processing of the selection process. If the selection probability of each user is directly calculated based on the information transfer degree of each user, and then the selection is performed, the selection probability of each user will be very small. The above content is to randomly select the information transfer degree, and then randomly select users within the range of similar information transfer degrees, splitting the original selection into two selections, so that the probability of each selection will not be too small, and the code can be written with a smaller data type.

[0115] Among them, the simplest solution to determine the selection probability based on the proportionality of the information transfer degree is to calculate the sum of all information transfer degrees, and then calculate the ratio of each information transfer degree to the sum, and the obtained ratio can be used as the selection probability.

[0116] Regarding step S400, the steps of obtaining the browsing content of the reference user, identifying the browsing content, and creating a keyword group include:

[0117] Query the browsing content of reference users within a preset time based on information acquisition permissions;

[0118] When the browsing content is audio information, performing audio recognition on it and converting it into text data;

[0119] When the browsing content is image information, the image information is recognized based on AI and converted into text data;

[0120] Extract keywords from text data to obtain keyword groups;

[0121] Compare and merge the keyword groups corresponding to all browsed contents, calculate the weight according to the number of groups corresponding to each keyword, and generate the weight item.

[0122] The above content details the process of generating keyword groups. The browsing content of reference users within a preset time is queried based on information acquisition permissions. The preset time generally adopts the time range preset by the management, such as one week. When the browsing content is audio information, audio recognition is performed on it and converted into text data. When the browsing content is image information, the image information is recognized based on AI and converted into text data. This process converts all data into text data. Audio recognition technology and AI are both existing technical means. It is worth mentioning that video data can be divided into audio data and image data.

[0123] By extracting keywords from text data, we can get keyword groups. Each text data can get a keyword group. All keyword groups are merged, the number of occurrences of each keyword is accumulated, and then the ratio of the number of occurrences to the total number of all occurrences is calculated as the weight.

[0124] Regarding step S500, when receiving news content containing a prediction request, the step of predicting the spread of the news content based on the keyword group includes:

[0125] When receiving news content containing a prediction request, extracting keywords of the news content based on the same recognition scheme as for browsing content;

[0126] Compare the extracted keywords with keyword groups to determine the degree of spread.

[0127] For news content that needs to be predicted, keywords of the news content are extracted based on the same recognition scheme as the browsing content, and compared with the constructed keyword groups to determine the degree of dissemination.

[0128] The calculation process of propagation degree is:

[0129] In the formula, S represents the propagation degree, B kis the weight of the keyword corresponding to the kth keyword in the keyword group; C4 is the preset correction coefficient; this process is very simple, which is to query the weight and then accumulate it. It is worth mentioning that when comparing keywords and keyword groups, synonyms can be regarded as the same word, which requires the use of existing text recognition technology, such as the bag-of-words model.

[0130] Figure 2 The structure diagram of the news content propagation prediction system based on big data is shown. In a preferred embodiment of the technical solution of the present invention, a news content propagation prediction system based on big data is also provided. The system 10 includes:

[0131] The information query module 11 is used to receive the information acquisition authority granted by the user, query the contacts of each user based on the information acquisition authority, and simultaneously query the activity and number of groups joined by the user;

[0132] The transfer degree calculation module 12 is used to determine the information transfer degree of the user according to the activity of the group the user joins and the number of people in the group;

[0133] A user selection module 13, configured to determine the selection probability of each user according to the information transfer degree, randomly select users based on the selection probability until the number of users reaches a preset threshold, and use the selected users as reference users;

[0134] The browsing content identification module 14 is used to obtain the browsing content of the reference user, identify the browsing content, and create a keyword group; the keyword group contains keyword items and weight items;

[0135] The spread degree prediction module 15 is used to predict the spread degree of the news content based on the keyword group when receiving news content containing a prediction request; the spread degree is used to characterize the spread ability of the news content, and the higher the spread degree, the stronger the spread ability.

[0136] Furthermore, the information query module 11 includes:

[0137] A contact query unit, used to receive information acquisition permissions granted by a user, and query contacts of each user based on the information acquisition permissions;

[0138] A group information acquisition unit locates the group in the contacts, obtains the number of group members and the total number of information in each time period;

[0139] An activity calculation unit, used to calculate the activity of the group according to the number of members in the group and the total number of messages in each time period;

[0140] The activity level is:

[0141] In the formula, H represents the activity, N represents the number of people in the group, N0 is the preset number of people, α, β and γ are the preset correction coefficients, and X max is the maximum number of messages in each period, X e is the average number of information in each period, T σ is the standard deviation of the total number of messages in each period.

[0142] Specifically, the transit degree calculation module 12 includes:

[0143] A time query unit, used to obtain the information sending time of the user in each group based on the information acquisition authority;

[0144] An affinity calculation unit, used for calculating the affinity between the user and any group according to the time when the information is sent;

[0145] The information transfer degree is calculated based on the user's affinity with each group and the activity of the corresponding group;

[0146] The affinity calculation process is as follows:

[0147]

[0148] The calculation process of the information transfer degree is:

[0149]

[0150] In the formula, Q represents affinity, N is the total number of information sending times, and t i represents the time when the i-th information is sent, X(t i ) represents the total number of messages in the time period corresponding to the i-th message sending time, X(t i-1 ) represents the total number of messages in the time period corresponding to the previous message sending time of the i-th message sending time, X(t i+1 ) represents the total number of messages in the time period corresponding to the next message sending time after the i-th message sending time; X(t0) and X(t N+1 ) is the default value; Z represents the information transfer degree, M represents the total number of user groups, and H j represents the activity of the jth group, Q j represents the affinity between the user and the jth group; C1, C2 and C3 are all preset constants; X(0) represents the preset number threshold.

[0151] In addition, the user selection module 13 includes:

[0152] A selection probability calculation unit, used to calculate the selection probability corresponding to the information transfer degree according to the direct proportion of the information transfer degree;

[0153] A loop selection unit, used for randomly selecting information transfer degrees based on a selection probability, cyclically selecting and recording the number of information transfer degrees, and exiting the loop when the number of information transfer degrees reaches a preset threshold;

[0154] The user query unit is used to determine a range based on the information transfer degree and a preset ratio, query users in the range, and randomly select a user as a reference user.

[0155] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A news content dissemination prediction method based on big data, characterized in that: The method comprises: Receive information access permissions granted by the user, query the contacts of each user based on the information access permissions, and simultaneously query the activity and number of groups that the user has joined; Determine the user's information transfer degree based on the activity of the group the user joins and the number of group members; Determine the selection probability of each user according to the information transfer degree, randomly select users based on the selection probability until the number of users reaches a preset threshold, and use the selected users as reference users; Obtain browsing content of a reference user, identify the browsing content, and create a keyword group; the keyword group contains keyword items and weight items; When news content containing a prediction request is received, the spread degree of the news content is predicted based on the keyword group; the spread degree is used to characterize the spread ability of the news content, and the higher the spread degree, the stronger the spread ability.

2. The news content dissemination prediction method based on big data according to claim 1 is characterized in that: The steps of receiving the information acquisition authority granted by the user, querying the contacts of each user based on the information acquisition authority, and simultaneously querying the activity and number of groups joined by the user include: Receive information access permissions granted by the user, and query the contacts of each user based on the information access permissions; Locate the group in the contacts, obtain the number of members in the group and the total number of messages in each time period; Calculate the activity of the group based on the number of people in the group and the total number of messages in each time period; The activity level is: In the formula, H represents the activity, N represents the number of people in the group, N0 is the preset number of people, α, β and γ are the preset correction coefficients, and X max is the maximum number of messages in each period, X e is the average number of information in each period, T σ is the standard deviation of the total number of messages in each period.

3. The news content dissemination prediction method based on big data according to claim 2 is characterized in that: The step of determining the user's information transfer degree according to the activity of the group the user joins and the number of people in the group includes: Obtain the time when a user sends information in each group based on the information acquisition authority; For any group, the affinity between the user and the group is calculated based on the time when the message was sent; The information transfer degree is calculated based on the user's affinity with each group and the activity of the corresponding group; The affinity calculation process is as follows: The calculation process of the information transfer degree is: In the formula, Q represents affinity, N is the total number of information sending times, and t i represents the time when the i-th information is sent, X(t i ) represents the total number of messages in the time period corresponding to the i-th message sending time, X(t i-1 ) represents the total number of messages in the time period corresponding to the previous message sending time of the i-th message sending time, X(t i+1 ) represents the total number of messages in the time period corresponding to the next message sending time after the i-th message sending time; X(t0) and X(t N+1 ) is the default value; Z represents the information transfer degree, M represents the total number of user groups, and H j represents the activity of the jth group, Q j represents the affinity between the user and the jth group; C1, C2 and C3 are all preset constants; X(0) represents the preset number threshold.

4. The news content dissemination prediction method based on big data according to claim 3 is characterized in that: The step of determining the selection probability of each user according to the information transfer degree, randomly selecting users based on the selection probability until the number of users reaches a preset threshold, and using the selected users as reference users includes: Calculating the selection probability corresponding to the information transfer degree according to the direct proportion of the information transfer degree; Randomly select information transfer degrees based on the selection probability, cyclically select and record the number of information transfer degrees, and exit the loop when the number of information transfer degrees reaches a preset threshold; A range is determined based on the information transfer degree and a preset ratio, and users in the range are queried, and a user is randomly selected as a reference user.

5. The news content dissemination prediction method based on big data according to claim 1 is characterized in that: The steps of obtaining browsing content of a reference user, identifying the browsing content, and creating a keyword group include: Query the browsing content of reference users within a preset time based on information acquisition permissions; When the browsing content is audio information, performing audio recognition on it and converting it into text data; When the browsing content is image information, the image information is recognized based on AI and converted into text data; Extract keywords from text data to obtain keyword groups; Compare and merge the keyword groups corresponding to all browsed contents, calculate the weight according to the number of groups corresponding to each keyword, and generate the weight item.

6. The news content dissemination prediction method based on big data according to claim 5 is characterized in that: When receiving news content containing a prediction request, the step of predicting the spread of the news content based on the keyword group includes: When receiving news content containing a prediction request, extracting keywords of the news content based on the same recognition scheme as for browsing content; Compare the extracted keywords with keyword groups to determine the degree of spread; The calculation process of propagation degree is: In the formula, S represents the propagation degree, B k is the weight of the keyword corresponding to the kth keyword in the keyword group; C4 is the preset correction coefficient.

7. A news content dissemination prediction system based on big data, characterized in that: The system comprises: The information query module is used to receive the information acquisition authority granted by the user, query the contacts of each user based on the information acquisition authority, and simultaneously query the activity and number of groups joined by the user; The transfer degree calculation module is used to determine the user's information transfer degree according to the activity of the group the user joins and the number of people in the group; A user selection module, used to determine the selection probability of each user according to the information transfer degree, randomly select users based on the selection probability until the number of users reaches a preset threshold, and use the selected users as reference users; A browsing content identification module is used to obtain the browsing content of the reference user, identify the browsing content, and create a keyword group; the keyword group contains keyword items and weight items; The spread degree prediction module is used to predict the spread degree of the news content based on the keyword group when receiving news content containing a prediction request; the spread degree is used to characterize the spread ability of the news content, and the higher the spread degree, the stronger the spread ability.

8. The news content dissemination prediction system based on big data according to claim 7 is characterized in that: The information query module includes: A contact query unit, used to receive information acquisition permissions granted by a user, and query contacts of each user based on the information acquisition permissions; A group information acquisition unit locates the group in the contacts, obtains the number of group members and the total number of information in each time period; An activity calculation unit, used to calculate the activity of the group according to the number of members in the group and the total number of messages in each time period; The activity level is: In the formula, H represents the activity, N represents the number of people in the group, N0 is the preset number of people, α, β and γ are the preset correction coefficients, and X max is the maximum number of messages in each period, X e is the average number of information in each period, T σ is the standard deviation of the total number of messages in each period.

9. The news content dissemination prediction system based on big data according to claim 8 is characterized in that: The transit degree calculation module comprises: A time query unit, used to obtain the information sending time of the user in each group based on the information acquisition authority; An affinity calculation unit, used for calculating the affinity between the user and any group according to the time when the information is sent; The information transfer degree is calculated based on the user's affinity with each group and the activity of the corresponding group; The affinity calculation process is as follows: The calculation process of the information transfer degree is: In the formula, Q represents affinity, N is the total number of information sending times, and t i represents the time when the i-th information is sent, X(t i ) represents the total number of messages in the time period corresponding to the i-th message sending time, X(t i-1 ) represents the total number of messages in the time period corresponding to the previous message sending time of the i-th message sending time, X(t i+1 ) represents the total number of messages in the time period corresponding to the next message sending time after the i-th message sending time; X(t0) and X(t N+1 ) is the default value; Z represents the information transfer degree, M represents the total number of user groups, and H j represents the activity of the jth group, Q j represents the affinity between the user and the jth group; C1, C2 and C3 are all preset constants; X(0) represents the preset number threshold.

10. The news content dissemination prediction system based on big data according to claim 9 is characterized in that: The user selection module comprises: A selection probability calculation unit, used to calculate the selection probability corresponding to the information transfer degree according to the direct proportion of the information transfer degree; A loop selection unit, used for randomly selecting information transfer degrees based on a selection probability, cyclically selecting and recording the number of information transfer degrees, and exiting the loop when the number of information transfer degrees reaches a preset threshold; The user query unit is used to determine a range based on the information transfer degree and a preset ratio, query users in the range, and randomly select a user as a reference user.

Images