CN110889526A - Method and system for predicting user upgrade complaint behavior - Google Patents
Method and system for predicting user upgrade complaint behavior Download PDFInfo
- Publication number
- CN110889526A CN110889526A CN201811041401.6A CN201811041401A CN110889526A CN 110889526 A CN110889526 A CN 110889526A CN 201811041401 A CN201811041401 A CN 201811041401A CN 110889526 A CN110889526 A CN 110889526A
- Authority
- CN
- China
- Prior art keywords
- sequence
- user
- event
- complaint
- preset
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0282—Rating or review of business operators or products
Abstract
The embodiment of the invention provides a method and a system for predicting a complaint behavior of a user upgrade, wherein the method comprises the following steps: the method comprises the steps of obtaining complaint behavior data of a user, and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data; extracting event items in the event sequence according to preset conditions, and constructing a time difference sequence of the complaint behaviors of the user; analyzing the time difference sequence of the complaint behaviors of the user as characteristics to obtain the probability of the upgrade complaint of the user corresponding to the time difference sequence of the complaint behaviors of the user; the preset conditions at least comprise a preset support threshold, a preset sequence maximum length and a preset time interval of an event item. According to the method provided by the invention, the original characteristics and the time sequence are combined to generate new characteristics by utilizing the user behavior sequence, and the user upgrading complaint behavior is predicted through the new characteristics, so that the dimension of the characteristics is improved, and the prediction precision of the user upgrading complaint behavior is improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of computers, in particular to a method and a system for predicting a complaint behavior of a user upgrading.
Background
With the development of mobile communication technology, mobile networks bring more convenience to people's lives. Meanwhile, the increase of the number of the users of the operators also makes the complaint problem of the customers increasingly prominent. Once the problem of the customer is not solved well, the customer will make further complaints, resulting in complaint upgrade, so that the acceptance of the customer on the brand of the operator is reduced, and on the other hand, complaints of upgrading the customer will also bring pressure to the assessment index of the operator.
In the prior art, more research directions are related predictions aiming at complaints of users, the further upgrade complaints of the users are not enough, the upgrade complaint predictions and the common complaint predictions are two completely different problems, and main factors influencing the complaints of the users, such as weak coverage of a wireless network, overhigh traffic charges, overhigh deductions and the like, are two main factors; the main factors influencing the customer upgrade complaints are the frequency of the customer complaints, the customer service solution, the time for solving the problem and the like.
On the other hand, complaint prediction technology for users is often performed by using a common classification algorithm, and because relevant features are considered statically, the complaint prediction method for upgrading the user often has poor performance, and the static features easily ignore the relationship between user behaviors and time variables.
Disclosure of Invention
The embodiment of the invention provides a method and a system for predicting a user upgrading complaint behavior, which are used for solving the problems that the complaint prediction technology for a user in the prior art is usually carried out by using a common classification algorithm, the performance of predicting the user upgrading complaint is often poor due to the fact that relevant characteristics are considered statically, and the relation between the user behavior and a time variable is easily ignored by the static characteristics.
In a first aspect, an embodiment of the present invention provides a method for predicting a complaint behavior of a user upgrade, including:
the method comprises the steps of obtaining complaint behavior data of a user, and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data;
extracting event items in the event sequence according to preset conditions, and constructing a time difference sequence of the complaint behaviors of the user;
analyzing the time difference sequence of the user complaint behavior as a characteristic to obtain the probability of the user for upgrading the complaint, which corresponds to the time difference sequence of the user complaint behavior;
the preset conditions at least comprise a preset support threshold, a preset sequence maximum length and a preset time interval of an event item.
In a second aspect, an embodiment of the present invention provides a system for predicting a complaint behavior of a user upgrade, including:
the device comprises a preprocessing module, a data processing module and a data processing module, wherein the preprocessing module is used for acquiring complaint behavior data of a user and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data;
the time difference sequence construction module is used for extracting event items in the event sequence according to preset conditions and constructing a time difference sequence of the complaint behaviors of the user;
the judging module is used for analyzing the time difference sequence of the user complaint behaviors as characteristics to obtain the probability of the user upgrading complaints corresponding to the time difference sequence of the user complaint behaviors;
the preset conditions at least comprise a preset support threshold, a preset sequence maximum length and a preset time interval of an event item.
In a third aspect, an embodiment of the present invention provides an electronic device, including a processor, a communication interface, a memory, and a bus, where the processor and the communication interface complete mutual communication through the bus, and the processor may call a logic instruction in the memory to execute the method for predicting a complaint behavior of a user upgrade according to the first aspect.
In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the method for predicting user-upgrade complaint behavior as described in the first aspect above.
According to the prediction method and system for the user upgrading complaint behaviors, the original features and the time sequence are combined to generate new features by the aid of the user behavior sequences, the user upgrading complaint behaviors are predicted through the new features, the behavior event sequence data of the users before and after a complaint order is generated by the users are utilized, frequent sequence pattern mining is conducted, accuracy of judging the user behaviors is improved by adding inter-item time constraints in the mining, feature dimensionality is improved, and prediction accuracy of upgrading complaint behaviors of the predicted users is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a method for predicting a complaint behavior of a user upgrade according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a method for predicting customer upgrade complaint behavior according to another embodiment of the present invention;
fig. 3 is a schematic structural diagram of a system for predicting customer upgrade complaint behavior according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flow chart of a method for predicting a complaint behavior of a user upgrade provided in an embodiment of the present invention, where the method includes:
s1, obtaining the complaint behavior data of the user, and establishing an event sequence for the user according to the event occurrence time and the event type in the complaint behavior data.
Specifically, the first choice needs to perform combing and data preparation on behavior data and service data related to the upgrade complaint, and the related data includes data of a broad complaint sheet and a narrow complaint sheet, data of a call contact, data of an upgrade call, data of a short message contact, data of a DPI, and the like. The data preparation result is to arrange the related events of the upgrade complaints in order according to the corresponding time, as shown in table 1, where table 1 is the related data of the upgrade complaints of a single user.
| User id | Time of day | Event(s) | Event symbol |
| 0000 | 2017/9/9 17:45 | Calling complaining telephone | a |
| 0000 | 2017/9/9 22:56 | Generating a complaint work order | b |
| 0000 | 2017/9/9 23:24 | Called complaining telephone | c |
| 0000 | 2017/9/9 23:31 | Unsatisfactory contact recovery | d |
| 0000 | 2017/9/10 12:58 | Complaint work order statement disapproval | e |
| 0000 | 2017/9/10 13:25 | Called complaining telephone | c |
| 0000 | 2017/9/10 13:52 | Dialing upgrade complaint related telephone | f |
TABLE 1
Constructing an event sequence according to the behaviors of a user in the process of complaint, and marking the time corresponding to the event, wherein the marked format is shown in the following formula:
Si＝<[t1]e1,[t2]e2,…,[ti]ei,…,[tn]en>
in the formula, SiA sequence of events representing a user i is also called a set of items, eiThe ith event representing the user, there being a term, tiRepresents an event eiThe time of occurrence, the resulting sequence of events is shown in table 2,
| user id | Sequence of events |
| 0000 | [2017/9/9 17:45]a,[2017/9/9 22:56]b,[2017/9/9 23:24]c,…… |
| 0001 | [2017/10/1 9:56]a,[2017/10/1 10:03]b,…… |
| …… |
TABLE 2
One for each action in the table, and a sequence of events for the user to act on the complaint.
By the method, the occurrence time of each behavior of the user is added into the event sequence, and support is made for the subsequent calculation of the time difference between different events.
S2, extracting event items in the event sequence according to preset conditions, and constructing a time difference sequence of the complaint behaviors of the user; the preset conditions at least comprise a preset support threshold, a preset sequence maximum length and a preset time interval of an event item.
Specifically, the parameters included in the preset conditions in this step are the preset support threshold α, the preset maximum sequence length L, and the preset time interval of the event item, after the event sequence of a single user is obtained in S1, for the event sequence, find out all prefixes with length 1 and corresponding projection databases, and mark the occurrence time t of the prefixes for suffixes in the projection databasesf1The difference from the original algorithm is mainly that the occurrence time of the suffix mark prefix in the projection database is increased to prepare for the subsequent calculation of the time difference between terms. When the time difference between two adjacent prefixes is not between the preset time intervals of the event items, if the time difference between two adjacent prefixes is smaller than the preset time interval, the user may repeat the complaint action in a short time, two identical actions may be regarded as the same action to be calculated, the item corresponding to one of the prefixes is removed, and if the time difference between two adjacent prefixes is greater than the preset time interval, the reason for the possible two complaints of the user is not the same reason, which may be regarded as two unrelated complaints, and at this time, the sequence may be classified into two different sequences.
Counting prefixes with the length of 1, deleting the items corresponding to the prefixes with the support degree lower than the threshold α from the data set S, and obtaining all frequent 1 item sequences, wherein i is 1. the support degree α threshold is mainly determined by the time consumed by the program, generally, the higher the support degree is, the more prefixes are deleted in each iteration, so that the subsequent useless calculation amount is reduced, and the overall calculation time is shorter, and conversely, the lower the support degree is, the less prefixes are deleted in each iteration, and the longer the overall calculation time is, the generated frequent sequences have the proportion in the positive and negative samples determined by the samples themselves.
And finally, judging whether the length of the prefix in the current event sequence is equal to the maximum length of the preset sequence, if so, taking the current event sequence as a time difference sequence of the complaint behaviors of the user, and if the length of the prefix in the current event sequence is smaller than the maximum length of the preset sequence, performing recursive mining on the prefix with the length of 1 and meeting the requirement of the support degree.
And S3, analyzing the time difference sequence of the user complaint behaviors as characteristics to obtain the probability of the user upgrading complaints corresponding to the time difference sequence of the user complaint behaviors.
Specifically, after the time difference sequence of the complaint behavior of the user is obtained, the time difference sequence is used as a feature to judge whether the complaint behavior of the user will be upgraded, and a specific implementation process can be that a mining mode of a frequent time difference sequence is used to screen the mode according to an expected hit rate threshold value in the sequence, or a neural network or a support vector machine is used to analyze the time difference sequence as a feature to obtain the probability that whether the user corresponding to the time difference sequence will be upgraded or not.
According to the method, the original characteristics and the time sequence are combined by using the user behavior sequence to generate new characteristics, the user upgrading complaint behavior is predicted through the new characteristics, the behavior event sequence data of the user before and after a complaint list is generated by the user, frequent sequence pattern mining is performed, the accuracy of judging the user behavior is improved by adding inter-item time constraints in the mining, the dimension of the characteristics is improved, and the prediction accuracy of upgrading the complaint behavior of the predicted user is improved.
On the basis of the above embodiment, the step of extracting event items in the event sequence according to preset conditions and constructing a time difference sequence of the complaint behavior of the user specifically includes:
in the event sequence, acquiring items with the length of 1 and a projection database corresponding to each item, and counting the occurrence time of the items for the item table with the length of 1; acquiring time intervals corresponding to every two adjacent items in the event sequence, and reserving the items of the time intervals between the preset time intervals of the event items; acquiring the support degrees of all items in the event sequence, and removing the items with the support degrees lower than a preset support degree threshold value to obtain a second event sequence; and counting the length of the second event sequence, and if the length of the second event sequence is within the maximum length of a preset sequence, taking the second event sequence as a time difference sequence of the complaint behaviors of the user.
Specifically, in the time calculation process of the event sequence, the projection database corresponding to the prefix of each item is found. Time of occurrence t of each suffix-labeled prefixfiCalculating the time difference between the prefix and its parent prefix, tf(i-1)-tfiAnd need to satisfy max gap>＝Δt>Min gap, terms that do not satisfy the condition will be discarded, rounding up the time difference Δ t (e.g., hourly granularity). If the projection database is empty, then a recursive return is made.
And after the time difference mining is carried out, counting the support degree counts of all items in the corresponding projection database. And removing items with the support degree lower than a preset support degree threshold value, and combining each single item meeting the support degree count with the current prefix to obtain a plurality of new prefixes. (if the time differences are different, the same sequence is different prefixes), a second event sequence is further constructed, the length of the second event sequence is counted, and if the length of the second event sequence is within the maximum length of the preset sequence, the second event sequence is used as a time difference sequence of the complaint behaviors of the user.
By the method, the logics of the front and back sequences of the user behaviors can be distinguished according to different time differences, so that frequent sequences belonging to different scenes are mined.
On the basis of the above embodiment, the step of analyzing the time difference sequence of the user complaint behavior as a feature to obtain the probability of the user performing the upgrade complaint corresponding to the time difference sequence of the user complaint behavior specifically includes:
and performing mode screening on the time difference sequence of the user complaint behaviors through a mining mode of the time frequent sequence to obtain the probability of the user upgrading complaints corresponding to the time difference sequence of the user complaint behaviors.
Specifically, in the step, a mining mode of the time difference frequent sequences is utilized, a group of time difference sequences can be found, the time difference sequences can be arranged from top to bottom according to the hit rate, and finally, mode screening is carried out according to an expected hit rate threshold.
In the concrete implementation, an event sequence is established for each complaint user, wherein complaints and upgrade complaints are positive samples, complaints and non-upgrade complaints are negative samples, and if the occurrence frequency of each item of a certain time difference sequence in the positive samples is nzThe number of occurrences in the negative example is nf. If the mode is met, the rule is judged to be 1, and the expected accuracy rate of the rule for predicting the upgrade complaint is as follows: auc ═ nz/(nz+nf). And screening the rule according to the expected accuracy rate, and realizing rule filtering on the production system.
The method provided by the embodiment has the advantage that the accuracy can reach a very high range through practical use. Because if a sequence of a pattern is elongated, such as a sequence of a pattern length of 10 for mining a frequent sequence, there is theoretically 10 regardless of the time difference between terms10And when various modes occur, the corresponding accuracy is different, and a batch of high-accuracy rules can be screened out to filter the user. Finally, the rules can be converted into 0-1 variables according to the existence of the rules, and the variables are input into the conventional model, so that the dimensionality of the data mining characteristics can be greatly expanded, and the accuracy of the model is improved.
On the basis of the above embodiment, the step of obtaining the support degrees of all the items in the event sequence, and removing the items whose support degrees are smaller than a preset support degree threshold value further includes: and if the support degrees of all the items in the event sequence are smaller than a preset support degree threshold value, judging that the user corresponding to the event sequence cannot carry out the complaint upgrading behavior.
Specifically, if the support counts of all the items in the event sequence are lower than the threshold α, it is determined that the user corresponding to the event sequence does not perform the upgrade complaint behavior, and the step is returned recursively.
On the basis of the foregoing embodiment, after the steps of obtaining the support degrees of all items in the event sequence, removing the items whose support degrees are smaller than the preset support degree threshold, and obtaining the second event sequence, the method further includes:
and merging the items with the support degree larger than a preset support degree threshold value, and re-executing the steps of extracting the event items in the event sequence according to preset conditions and constructing the time difference sequence of the complaint behaviors of the user according to the second event sequence constructed by merging.
Specifically, each single item meeting the support degree count is merged with the current prefix to obtain a plurality of new prefixes. (if the time differences are different, the same sequence is a different prefix) to obtain a second event sequence, and recursively executing the steps on the second event sequence until the sequence length meets the preset requirement.
On the basis of the foregoing embodiment, the step of counting the length of the second event sequence, and if the length of the second event sequence is within a preset maximum sequence length, taking the second event sequence as a time difference sequence of the complaint behavior of the user further includes: and when the length of the second event sequence is greater than the maximum length of the preset sequence, extracting a part of the second event sequence with the length equal to the maximum length of the preset sequence as a time difference sequence of the complaint behaviors of the user.
The step of counting the length of the second event sequence, and if the length of the second event sequence is within the maximum length of the preset sequence, taking the second event sequence as the time difference sequence of the complaint behavior of the user further includes: and if the length of the second event sequence is smaller than the maximum length of the preset sequence, taking the second event sequence as a time difference sequence of the complaint behaviors of the user.
Specifically, when the second event length is greater than a preset maximum sequence length L, a part of the second event sequence with a length equal to the maximum length of the preset sequence is extracted, for example, when the maximum sequence length L is set to 10, when the number of terms in the second event sequence exceeds 10 terms, the first 10 terms in the second event sequence are extracted, and a sequence composed of the first 10 terms is used as a time difference sequence of the complaint behavior of the user. And if the length of the constructed second event sequence is smaller than the maximum sequence length L, directly performing the characteristic judgment step by taking the second event sequence as the time difference sequence of the complaint behavior of the user.
In summary, according to the method for predicting the user upgrade complaint behavior provided by the embodiment of the present invention, the original feature and the time sequence are combined by using the user behavior sequence to generate a new feature, and the user upgrade complaint behavior is predicted by using the new feature. Because the new features are obtained based on the original features and the time series, the new features comprise other features which are different from the original features and are more valuable for prediction, and the prediction result of the embodiment is more accurate.
In another embodiment of the present invention, referring to fig. 2, fig. 2 is a schematic flowchart of a method for predicting a complaint behavior of a user upgrade according to another embodiment of the present invention, where the method includes the following specific steps:
202, the data is pre-processed to convert the data into a time difference sequence format, as shown in table 2 above.
203 parameters set that the inputs 202 produce are the time difference sequence S, the support threshold α, the maximum sequence length L, the maximum inter-term interval max gap, and the minimum interval min gap.
204 find out all prefixes with the length of 1 and corresponding projection databases in the time difference sequence S, and mark the occurrence time t of the prefixes with the suffixes in the projection databasesf1。
205, counting prefixes with length 1 in the projection database generated in 204, deleting the items corresponding to prefixes with support degree lower than threshold α from the data set S, and obtaining all the frequent 1 item sequences, i ═ 1
206 judging whether the length reaches the set maximum sequence length L, if so, directly jumping to a post-processing link. If not, recursive mining is required for each prefix of length i that meets the support requirement.
207 find the currentAnd the projection database corresponding to the prefix. Time of occurrence t of each suffix-labeled prefixfiThe time difference is calculated to satisfy the foregoing condition of Δ t, and the time difference Δ t is rounded up (e.g., hourly granularity). If the time differences are not the same, the same sequence is a different prefix. And will return recursively if the projection database is empty.
208 count the support counts for the entries in the corresponding projection database, if the support counts for all entries are below the threshold α, then a recursive return is made.
209 merge the individual items satisfying the support count with the current prefix to get several new prefixes.
210, let i be i +1, the prefixes are the prefixes after merging the single entries, and jump to 206.
Fig. 3 is a schematic structural diagram of a customer upgrade complaint behavior prediction system according to an embodiment of the present invention, and as shown in fig. 3, the provided system includes: a preprocessing module 31, a time difference sequence constructing module 32 and a judging module 33.
The preprocessing module 31 is configured to obtain complaint behavior data of a user, and establish an event sequence for the user according to event occurrence time and event type in the complaint behavior data;
the time difference sequence construction module 32 is configured to extract event items in the event sequence according to preset conditions, and construct a time difference sequence of the complaint behaviors of the user;
the judging module 33 is configured to analyze the time difference sequence of the user complaint behavior as a feature to obtain a probability that the user corresponding to the time difference sequence of the user complaint behavior carries out the upgrade complaint;
the preset conditions at least comprise a preset support threshold, a preset sequence maximum length and a preset time interval of an event item.
It should be noted that, the preprocessing module 31, the time difference sequence constructing module 32, and the determining module 33 cooperate to execute the method for predicting the user upgrade complaint behavior in the foregoing embodiment, and specific functions of the system refer to the foregoing embodiment of the method for predicting the user upgrade complaint behavior, which is not described herein again.
Fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 4, a vulnerability detection device of a deep learning system includes: a processor (processor)401, a communication Interface (communication Interface)402, a memory (memory)403, and a bus 404, wherein the processor 401 and the memory 403 communicate with each other through the bus 404. Processor 401 may call logic instructions in memory 403 to perform methods including, for example: the method comprises the steps of obtaining complaint behavior data of a user, and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data; extracting event items in the event sequence according to preset conditions, and constructing a time difference sequence of the complaint behaviors of the user; and analyzing the time difference sequence of the user complaint behaviors as characteristics to obtain the probability of the user for upgrading complaints corresponding to the time difference sequence of the user complaint behaviors.
An embodiment of the present invention discloses a computer program product, which includes a computer program stored on a non-transitory computer readable storage medium, where the computer program includes program instructions, and when the program instructions are executed by a computer, the computer can execute the method provided by the above method embodiments, for example, the method includes: the method comprises the steps of obtaining complaint behavior data of a user, and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data; extracting event items in the event sequence according to preset conditions, and constructing a time difference sequence of the complaint behaviors of the user; and analyzing the time difference sequence of the user complaint behaviors as characteristics to obtain the probability of the user for upgrading complaints corresponding to the time difference sequence of the user complaint behaviors.
The present embodiments provide a non-transitory computer-readable storage medium storing computer instructions that cause a computer to perform the methods provided by the above method embodiments, for example, including: the method comprises the steps of obtaining complaint behavior data of a user, and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data; extracting event items in the event sequence according to preset conditions, and constructing a time difference sequence of the complaint behaviors of the user; and analyzing the time difference sequence of the user complaint behaviors as characteristics to obtain the probability of the user for upgrading complaints corresponding to the time difference sequence of the user complaint behaviors.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for predicting a complaint behavior of a user upgrade is characterized by comprising the following steps:
the method comprises the steps of obtaining complaint behavior data of a user, and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data;
extracting event items in the event sequence according to preset conditions, and constructing a time difference sequence of the complaint behaviors of the user;
analyzing the time difference sequence of the user complaint behavior as a characteristic to obtain the probability of the user for upgrading the complaint, which corresponds to the time difference sequence of the user complaint behavior;
the preset conditions at least comprise a preset support threshold, a preset sequence maximum length and a preset time interval of an event item.
2. The method according to claim 1, wherein the step of extracting event items in the event sequence according to preset conditions and constructing a time difference sequence of the complaint behavior of the user specifically comprises:
in the event sequence, acquiring items with the length of 1 and a projection database corresponding to each item, and counting the occurrence time of the items for the item table with the length of 1;
acquiring time intervals corresponding to every two adjacent items in the event sequence, and reserving the items of the time intervals between the preset time intervals of the event items;
acquiring the support degrees of all items in the event sequence, and removing the items with the support degrees lower than a preset support degree threshold value to obtain a second event sequence;
and counting the length of the second event sequence, and if the length of the second event sequence is within the maximum length of a preset sequence, taking the second event sequence as a time difference sequence of the complaint behaviors of the user.
3. The method according to claim 1, wherein the step of analyzing the time difference sequence of the user complaint behavior as a feature to obtain a probability of the user complaint being upgraded corresponding to the time difference sequence of the user complaint behavior specifically comprises:
and performing mode screening on the time difference sequence of the user complaint behaviors through a mining mode of the time frequent sequence to obtain the probability of the user upgrading complaints corresponding to the time difference sequence of the user complaint behaviors.
4. The method according to claim 2, wherein the step of obtaining the support of all the items in the event sequence and removing the items with the support smaller than a preset support threshold further comprises:
and if the support degrees of all the items in the event sequence are smaller than a preset support degree threshold value, judging that the user corresponding to the event sequence cannot carry out the complaint upgrading behavior.
5. The method according to claim 2, wherein the step of obtaining the support degrees of all the items in the event sequence, removing the items with the support degrees smaller than a preset support degree threshold, and obtaining the second event sequence further comprises:
and merging the items with the support degree larger than a preset support degree threshold value, and re-executing the steps of extracting the event items in the event sequence according to preset conditions and constructing the time difference sequence of the complaint behaviors of the user according to the second event sequence constructed by merging.
6. The method of claim 2, wherein the step of counting the length of the second event sequence and regarding the second event sequence as a time difference sequence of the complaint behavior of the user if the length of the second event sequence is within a preset maximum sequence length further comprises:
and when the length of the second event sequence is greater than the maximum length of the preset sequence, extracting a part of the second event sequence with the length equal to the maximum length of the preset sequence as a time difference sequence of the complaint behaviors of the user.
7. The method of claim 2, wherein the step of counting the length of the second event sequence and regarding the second event sequence as a time difference sequence of the complaint behavior of the user if the length of the second event sequence is within a preset sequence maximum length further comprises:
and if the length of the second event sequence is smaller than the maximum length of the preset sequence, taking the second event sequence as a time difference sequence of the complaint behaviors of the user.
8. A customer escalation complaint behavior prediction system, comprising:
the device comprises a preprocessing module, a data processing module and a data processing module, wherein the preprocessing module is used for acquiring complaint behavior data of a user and establishing an event sequence for the user according to event occurrence time and event types in the complaint behavior data;
the time difference sequence construction module is used for extracting event items in the event sequence according to preset conditions and constructing a time difference sequence of the complaint behaviors of the user;
the judging module is used for analyzing the time difference sequence of the user complaint behaviors as characteristics to obtain the probability of the user upgrading complaints corresponding to the time difference sequence of the user complaint behaviors;
the preset conditions at least comprise a preset support threshold, a preset sequence maximum length and a preset time interval of an event item.
9. An electronic device, comprising a processor, a communication interface, a memory and a bus, wherein the processor, the communication interface and the memory communicate with each other via the bus, and the processor can call logic instructions in the memory to execute the method for predicting customer-updated complaint behavior according to any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the method for predicting customer escalation complaint behavior according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811041401.6A CN110889526B (en) | 2018-09-07 | 2018-09-07 | User upgrade complaint behavior prediction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811041401.6A CN110889526B (en) | 2018-09-07 | 2018-09-07 | User upgrade complaint behavior prediction method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110889526A true CN110889526A (en) | 2020-03-17 |
| CN110889526B CN110889526B (en) | 2022-06-28 |
Family
ID=69744433
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811041401.6A Active CN110889526B (en) | 2018-09-07 | 2018-09-07 | User upgrade complaint behavior prediction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110889526B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563754A (en) * | 2020-04-29 | 2020-08-21 | 云南电网有限责任公司信息中心 | Service sequence driven power customer appeal perception system |
| CN112819245A (en) * | 2021-02-24 | 2021-05-18 | 阳光保险集团股份有限公司 | User complaint behavior prediction method, device, equipment and storage medium |
| CN113660380A (en) * | 2021-08-16 | 2021-11-16 | 西安京迅递供应链科技有限公司 | Information processing method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7493366B1 (en) * | 2001-03-01 | 2009-02-17 | Verizon Laboratories, Inc. | System and method for processing customer requests relating to unsolicited commercial email and other service disruptions |
| CN104618949A (en) * | 2015-02-13 | 2015-05-13 | 浪潮通信信息系统有限公司 | Complaint predicting method and device based on ARMA model |
| CN106203631A (en) * | 2016-07-05 | 2016-12-07 | 中国科学院计算技术研究所 | The parallel Frequent Episodes Mining of description type various dimensions sequence of events and system |
| CN106685674A (en) * | 2015-11-05 | 2017-05-17 | 华为技术有限公司 | Network event prediction and method and apparatus for establishing network event prediction model |
| CN107992609A (en) * | 2017-12-15 | 2018-05-04 | 广东电网有限责任公司信息中心 | A kind of complaint tendency determination methods based on Text Classification and decision tree |
-
2018
- 2018-09-07 CN CN201811041401.6A patent/CN110889526B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7493366B1 (en) * | 2001-03-01 | 2009-02-17 | Verizon Laboratories, Inc. | System and method for processing customer requests relating to unsolicited commercial email and other service disruptions |
| CN104618949A (en) * | 2015-02-13 | 2015-05-13 | 浪潮通信信息系统有限公司 | Complaint predicting method and device based on ARMA model |
| CN106685674A (en) * | 2015-11-05 | 2017-05-17 | 华为技术有限公司 | Network event prediction and method and apparatus for establishing network event prediction model |
| CN106203631A (en) * | 2016-07-05 | 2016-12-07 | 中国科学院计算技术研究所 | The parallel Frequent Episodes Mining of description type various dimensions sequence of events and system |
| CN107992609A (en) * | 2017-12-15 | 2018-05-04 | 广东电网有限责任公司信息中心 | A kind of complaint tendency determination methods based on Text Classification and decision tree |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563754A (en) * | 2020-04-29 | 2020-08-21 | 云南电网有限责任公司信息中心 | Service sequence driven power customer appeal perception system |
| CN112819245A (en) * | 2021-02-24 | 2021-05-18 | 阳光保险集团股份有限公司 | User complaint behavior prediction method, device, equipment and storage medium |
| CN113660380A (en) * | 2021-08-16 | 2021-11-16 | 西安京迅递供应链科技有限公司 | Information processing method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110889526B (en) | 2022-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8861691B1 (en) | Methods for managing telecommunication service and devices thereof | |
| CN107566358B (en) | Risk early warning prompting method, device, medium and equipment | |
| CN110889526B (en) | User upgrade complaint behavior prediction method and system | |
| CN102083010B (en) | Method and equipment for screening user information | |
| US20160232452A1 (en) | Method and device for recognizing spam short messages | |
| CN111294819B (en) | Network optimization method and device | |
| CN107645740A (en) | A kind of mobile monitoring method and terminal | |
| CN105824818A (en) | Informationized management method, platform and system | |
| CN115470867A (en) | Agent matching method, device, equipment and storage medium based on knowledge graph | |
| US9398441B2 (en) | Method and apparatus for identifying re-subscribed user | |
| CN109428760B (en) | User credit evaluation method based on operator data | |
| CN108076032B (en) | Abnormal behavior user identification method and device | |
| CN110677269B (en) | Method and device for determining communication user relationship and computer readable storage medium | |
| US8594623B2 (en) | Subscriber portfolio management system | |
| US20170034035A1 (en) | Real time event monitoring and analysis system | |
| CN114218500B (en) | User mining method, system, device and storage medium | |
| CN110582091A (en) | method and apparatus for locating wireless quality problems | |
| CN114881521A (en) | Service evaluation method, device, electronic equipment and storage medium | |
| Nadaf et al. | Data mining in telecommunication | |
| CN110597572B (en) | Service call relation analysis method and computer system | |
| CN110087230B (en) | Data processing method, data processing device, storage medium and electronic equipment | |
| CN112308419A (en) | Data processing method, device, equipment and computer storage medium | |
| Badarch et al. | Big data area: A novel network performance analysis technique based on bayesian traffic classification algorithm | |
| CN115529566B (en) | Charging control method, device and storage medium based on predefined Urr | |
| Shi et al. | Segmentation of mobile user groups based on traffic usage and mobility patterns |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |