Time-series and time-slot based user behavior pattern mining method and system

Abstract

The application provides a time sequence and time slot-based user behavior pattern mining method and system, and relates to the technical field of user behavior pattern mining.The method comprises the following steps: arranging user behavior information in a daily routine log in a sequence according to the order of time to obtain a user behavior sequence; constructing a suffix database containing time intervals based on the user behavior sequence; introducing a time factor when establishing the suffix database; adding a time interval between a prefix and a next item set to each suffix corresponding to the prefix; determining that the behavior in the frequent pattern is allowed to be fuzzy in sequence within the time interval; comprehensively considering the time sequence and time interval information between behaviors to mine a user behavior frequent pattern; and using the frequent pattern based on the time sequence and time slot to represent the living habits and living rules of a user.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of user behavior analysis, in particular to a user behavior pattern mining method and system based on time sequence and time slot. BACKGROUND

[0002] With the rapid development of technology, people have higher requirements for the quality of life, and pursue a healthy and intelligent lifestyle. More and more users use smart bracelets, smart home devices and other devices to monitor daily living behavior logs, find daily behavior frequent patterns from daily living logs, find out user's living habits and living rules, and evaluate the health status of life. Therefore, it is particularly important to mine user's daily behavior frequent patterns from daily living logs. Usually, daily living logs include user's daily behavior and the corresponding time of the behavior, and the traditional mining algorithm only considers the relevance of user behavior, and the frequent patterns mined are difficult to accurately represent user's behavior habits and living rules. The behavior representation form construction strategy based on time sequence and time slot not only focuses on the relevance of behavior, but also considers the important factors of behavior sequence and time interval, and the frequent patterns finally mined can represent the time sequence of user's living habits and the time interval, and realize the maximum mining of valuable and meaningful information in user's daily living logs, better representing user's behavior habits and living rules.

[0003] Frequent pattern mining is the core problem of user behavior analysis. Frequent pattern is a pattern that frequently appears in a sample data set. "Pattern" is a relatively abstract concept word. For example, in a daily living log within a certain period of time, the behavior log corresponding to the time in each day is recorded, and each record information includes the time of behavior occurrence and the corresponding behavior. If these data are analyzed and mined, it is found that eating dinner and watching TV always appear in succession in the records of a day, and the frequency reaches a specified threshold. Then, the two behaviors of eating dinner and watching TV appearing in the records of the same day are called a frequent pattern, and such a data mining algorithm is called frequent pattern mining. Commonly used frequent pattern mining algorithms include Apriori, Prefixspan, etc.

[0004] Apriori algorithm is a classic algorithm in data mining, and its core is to find candidate item sets by repeatedly traversing the database and filter out frequent item sets through support, to mine frequent patterns in an iterative manner. Although the Apriori algorithm can mine the relevance between user behaviors from a large number of daily living logs, it cannot find the time sequence and time interval information between behaviors.

[0005] Prefixspan algorithm is full name of Prefix-Projected Pattern Growth, namely prefix projected pattern mining. The core of the algorithm adopts the idea of divide and conquer, constantly generates multiple smaller projected databases of the sequence database, and then performs sequence pattern mining on each projected database. The Prefixspan algorithm can not only mine the correlation of user behavior in user daily behavior log, but also further mine the timing information of the behavior, but still ignores the important time factor in daily routine log.

[0006] The existing frequent pattern mining algorithms such as Apriori, Prefixspan and the like only consider the mining of behavior correlation or timing, and none of them comprehensively considers the timing and time interval information between behaviors from both behavior factor and time factor. The life habits and life rules of users are derived from daily routine log, and the daily routine log is usually composed of two elements of behavior and corresponding time. It is difficult to find life rules containing rich information and having significance from the perspective of behavior correlation or timing. Therefore, how to comprehensively consider the timing and time interval information between behaviors from both behavior factor and time factor, mine meaningful frequent behavior patterns from user daily routine log, and represent life habits and life rules of users is a problem to be solved. SUMMARY

[0007] In view of this, the application provides a user behavior pattern mining method based on timing and time slot, solves the problem that the traditional frequent pattern mining technology can only mine frequent patterns containing behavior information in user behavior pattern mining, and ignores time information, so as to realize mining of frequently occurring patterns containing time information and behavior information from a large number of daily routine logs, and representing life habits and life rules of users.

[0008] To this end, the application provides the following technical solutions:

[0009] The application provides a user behavior pattern mining method based on timing and time slot, which comprises the following steps:

[0010] A suffix database containing time interval is constructed based on user behavior sequence. When the suffix database is established, a time interval between a prefix and a next item set is introduced into each suffix corresponding to the prefix.

[0011] Based on the suffix database, the timing and time interval information between behaviors are comprehensively considered by determining that the behaviors in the frequent pattern are allowed to be fuzzy in order within the time interval, and the user behavior frequent pattern is mined.

[0012] The frequent pattern based on timing and time slot represents the life habits and life rules of users.

[0013] Further, considering the time sequence and time interval information between behaviors, the user behavior frequent pattern is mined, including:

[0014] S1, scanning the user daily routine log database to obtain a candidate item set, and screening a frequent item set from the candidate item set as a frequent k-sequence set according to a set support degree, k representing a sequence length in the sequence set;

[0015] S2, taking each frequent 1-sequence in the frequent 1-sequence set as a prefix, and establishing a time interval containing suffix database corresponding to each frequent 1-sequence through the time interval containing suffix database;

[0016] S3, scanning the above-mentioned constructed time interval containing suffix database, screening a frequent item set as a suffix of the current frequent sequence according to a set support degree, and performing a connection operation with the corresponding prefix to form a frequent k+1-sequence set;

[0017] S4, taking each frequent k+1-sequence in the new frequent k+1-sequence set as a prefix, and establishing a time interval containing suffix database corresponding to each frequent k+1-sequence through the time interval containing suffix database;

[0018] Repeating the above-mentioned S3 and S4, and recursively mining until all frequent sequence patterns in the daily routine log database are obtained.

[0019] Further, the time interval containing suffix database is constructed, including:

[0020] The starting time sequence corresponding to the user behavior sequence is obtained from the daily routine log, and is represented as follows:

[0021] S T ={t1,t2,...t n};

[0022] Wherein, t i represents the starting time corresponding to the i-th user behavior in S A , and n represents n user behaviors in S A ; S A ={a1,a2,...a n}, a i ∈{a,z}; S A represents a user behavior sequence, each lowercase letter represents a user behavior, and the same user behavior is represented by the same letter, and n represents n user behaviors contained in S A ;

[0023] According to the user behavior sequence S A and the corresponding time sequence ST Construct a sequence of triples to store user actions, corresponding times, and time intervals between adjacent actions, as shown below:

[0024] S I ={(a2,t2,I2),(a3,t3,I3)...(a n ,t n ,I n )};

[0025] Among them, I i =t i -t i-1 n represents S I There are a total of n-1 triplet elements;

[0026] Given a sequence of triplets, obtain the suffix sequence with the i-th tuple as the prefix, as shown below:

[0027] S′ I ={(a i+1 ,t i+1 ,I i+1 ),(a i+2 ,t i+2 ,I i+2 ),...(a n ,t n ,I n )};

[0028] Where n represents S′ I There are ni elements in total;

[0029] Obtain the tuple (a) based on the suffix sequence. i ,t i ,I i The time interval suffix database with a prefix of ) is represented as follows:

[0030] DataBase = {D1, D2, ... D} m};

[0031] Among them, D k Represents a tuple (a i ,t i ,I i The k-th suffix sequence S′ with prefix ) I m represents the tuple (a i ,t i ,I i The time interval suffix database with ) as the prefix contains a total of m suffix sequences.

[0032] Furthermore, users' lifestyle habits and routines are characterized by frequent patterns based on time series and time slots, including:

[0033] Each behavior in the time-series and time-slot based frequent pattern sequence set represents a user's behavior frequent pattern alternated with time, and is expressed as follows:

[0034] S={a1, I1, a2, I2, …, a n-1 , I n-1 , a n};

[0035] Wherein, a i is a certain letter in a-z, representing a user's behavior; I i is a non-negative integer, representing the interval time between a i and a i+1 . With such a time-series and time-slot based frequent pattern sequence, a user's living habit and living rule are represented.

[0036] The application also provides a time-series and time-slot based user behavior pattern mining system, which comprises:

[0037] A user behavior sequence unit, which arranges the user behavior information in the daily routine log in sequence according to the order, and obtains a user behavior sequence;

[0038] A suffix database establishment unit, which constructs a suffix database containing time intervals based on the user behavior sequence obtained by the user behavior sequence unit, and introduces time factor in the establishment of the suffix database, and adds a time interval between the prefix and the next item set into each suffix corresponding to the prefix;

[0039] A frequent pattern mining unit, which mines the user behavior frequent pattern based on the suffix database established by the suffix database establishment unit, and determines the order ambiguity of the behavior in the frequent pattern within the time interval, and comprehensively considers the time sequence and time interval information between the behaviors;

[0040] A user behavior pattern representation unit, which represents the user's living habit and living rule with the time-series and time-slot based frequent pattern obtained by the frequent pattern mining unit.

[0041] The present application has the advantages and positive effects that the present application introduces time factor when establishing suffix database, adds a time interval between a prefix and the next item set into each suffix corresponding to the prefix, and constructs an extended suffix database containing time interval, and defines the time interval as time slot. The present application constructs a suffix database containing time interval when mining frequent pattern, represents user's living habit and living rule based on time sequence and time slot information of frequent pattern, and allows the behaviors in frequent pattern to be fuzzy in sequence within the determined time interval, which is in accordance with the actual situation of daily life log. The user representation form construction strategy based on time sequence and time slot comprehensively considers the time sequence between behaviors and the time interval information thereof from two aspects of behavior factor and time factor, mines meaningful frequent behavior pattern from daily life log of user, and represents user's living habit and living rule. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0043] Figure 1 The flow chart of the user behavior pattern mining method based on time sequence and time slot in the embodiments of the present application;

[0044] Figure 2 The flow chart of the user behavior pattern mining method based on daily life log in the embodiments of the present application;

[0045] Figure 3 The flow chart of the construction method of time interval suffix database in the embodiments of the present application. DETAILED DESCRIPTION

[0046] In order to make the technical personnel of the present application better understand the technical solutions, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0047] It is to be understood that the terminology "first", "second", and the like used in the specification and the claims of the application as well as the foregoing drawings is merely used to distinguish similar objects and is not to be construed as a specific order or as having a particular precedence over one another. It should be understood that the use of such terminology herein is only intended to ease the description of the embodiments of the application described herein and is not intended to limit the scope of the application as claimed. Furthermore, the terms "comprise" and "comprising" and any variations thereof as used in the specification and the claims of the application are intended to cover a non-exclusive inclusion, such that processes, methods, articles, or apparatuses that comprise a list of steps or elements, without necessarily being limited to the steps or elements that are expressly listed in the list, are still considered to be within the scope of the present application.

[0048] The application proposes a time interval and time slot based user behavior pattern mining method, introduces time factor when establishing suffix database, adds a time interval between prefix and next item set into each suffix corresponding to the prefix to construct suffix database containing time interval; determines that there is ambiguity in sequence of behavior in frequent pattern in time interval, comprehensively considers time sequence between behaviors and time interval information, and mines user behavior frequent pattern; and the frequent pattern based on time interval and time slot represents user's living habit and living rule.

[0049] As shown in Figure 1 The application proposes a time interval and time slot based behavior representation form and construction strategy including three main stages of constructing user behavior database, obtaining frequent item set and generating new frequent sequence, and constructing time interval suffix database.

[0050] The time interval and time slot based user behavior pattern mining method based on daily living log is described in detail below with specific examples.

[0051] The data set is described as follows:

[0052] The daily living log data set contains 21 days of user living log;

[0053] In time, each day starts from 0:00 and ends at 24:00;

[0054] In behavior, a total of ten daily behavior labels are included, which are: Leaving, Toileting, Showering, Sleeping, Breakfast, Lunch, Dinner, Snack, Spare_Time / TV, and Grooming.

[0055] Task: by constructing time interval suffix database, mining user behavior frequent pattern containing time interval and time slot, representing user's living habit and living rule.

[0056] The application introduces time factor when establishing projection database, constructs suffix database containing time interval, comprehensively considers behavior factor and time factor to mine frequent pattern, discovers user behavior frequent pattern containing time sequence and time slot, represents user living habit and living rule, and the whole method flow is as shown in Figure 2 Figure 2 In the formula, min_sup is support threshold value, is non-negative number not more than 1, represents that user behavior pattern frequency exceeds min_sup, and is frequent behavior pattern; t is time interval value, represents that t length is allowed to exist in user behavior order fuzzy; ItemSet represents newly generated frequent item set. In the embodiment, min_sup is set to 0.4, and t is set to 1.

[0057] The user behavior frequent pattern mining method considering time sequence between behaviors and time interval information is as follows:

[0058] S1, scanning user daily living log database to obtain candidate item set, screening frequent item set from the candidate item set according to the set support, taking the frequent item set as frequent k-sequence set, k represents sequence length in sequence set, and specifically, the user behavior number contained in the sequence, and at this time, the k value is 1;

[0059] S2, taking each frequent 1-sequence in the frequent 1-sequence set as a prefix, establishing the respective time interval containing suffix database of each frequent 1-sequence through the time interval containing suffix database;

[0060] S3, scanning the above-mentioned time interval containing suffix database, screening frequent item set according to the set support, taking the frequent item set as the suffix of the current frequent sequence, and performing connection operation with the corresponding prefix, to form frequent (k+1) sequence set;

[0061] S4, taking each frequent (k+1) sequence in the new frequent (k+1) sequence set as a prefix, establishing the respective time interval containing suffix database of each frequent (k+1) sequence through the time interval containing suffix database;

[0062] Repeating the above-mentioned S3 and S4, recursively mining until all frequent sequence patterns in the daily living log database are obtained.

[0063] The part of user living habit mined based on the daily living log time factor and behavior factor is shown in table 1, and in table 1, lowercase letters a-j represent the following ten behavior activities: Leaving, Toileting, Showering, Sleeping, Breakfast, Lunch, Dinner, Snack, Spare_Time / TV, Grooming;​k represents the time interval between k and k+1 hours.

[0064] Table 1

[0065]

[0066] The results of Table 1 show part of the user behavior patterns mined based on the time factor and the behavior factor of the daily life log. For example, the sequence No. 5 sequence represents that the user has the following living habits for 10 days out of 21 days: getting up to have breakfast after sleeping for 8-9 hours, having lunch after an interval of 5-6 hours, and having dinner again after an interval of 5-6 hours. From the sequence No. 5 sequence, it can be known that the user usually sleeps for 8-9 hours, and has regular three meals, and has a relatively healthy life state. For example, the sequence No. 1 sequence represents that the user has the following living habits for 17 days out of 21 days: having some beauty and skin care after leisure and entertainment for 0-1 hours, and going to sleep after an interval of 0-1 hours. From the sequence No. 1 sequence, it can be known that the user often has some leisure and entertainment and beauty and skin care activities before sleeping, and such results enable the smart home system to understand the living habits of the user more, and thus provide more intelligent services.

[0067] The present application introduces the time factor when establishing the suffix database, as shown in the following formula: Figure 3 The construction method of the time interval suffix database is as follows:

[0068] The user behavior information in the daily life log of each day is arranged in sequence according to the order, to obtain a user behavior sequence, which is represented as shown in the following formula:

[0069] S A ={a1,a2,...a n},a i ∈{a,z};

[0070] wherein each lowercase letter represents a user behavior, the same user behavior is represented by the same letter, and n represents that S A contains n user behaviors in total.

[0071] The starting time sequence corresponding to the user behavior sequence is obtained from the daily life log, which is represented as shown in the following formula:

[0072] S T ={t1,t2,...t n};

[0073] wherein t i represents the starting time corresponding to the i th user behavior in S A , and n represents that S A contains n user behaviors in total.

[0074] According to the user behavior sequence S Aand the corresponding time sequence S T Construct a triple sequence, respectively, store user behavior, corresponding time, time interval of adjacent behavior, as shown below:

[0075] S I = {(a2, t2, I2), (a3, t3, I3)... (an, tn, In)}; n n n

[0076] Where I i = t i - t i-1 , n represents n-1 triple elements in S I .

[0077] According to the triple sequence, get the suffix sequence prefixed with the i-th tuple, as shown below:

[0078] S' I = {(a i+1 , t i+1 , I i+1 ), (a i+2 , t i+2 , I i+2 ),... (a n , t n , I n )};

[0079] Where n represents n-i elements in S' I .

[0080] According to the suffix sequence, get the time interval suffix database prefixed with the tuple (a i , t i , I i ), as shown below:

[0081] DataBase = {D1, D2,...D m};

[0082] Where D k represents the k-th suffix sequence S' I prefixed with the tuple (a i , t i , I i ), and m represents the time interval suffix database prefixed with the tuple (a i , t i , I i ) contains m suffix sequences.

[0083] The method of representing user's life habits and life rules based on time sequence and time slot is as follows:​​​

[0084] After the above user behavior frequent pattern mining, a frequent pattern sequence set composed of behavior and time interval alternately can be obtained, wherein each behavior and time alternately sequence represents a user behavior frequent pattern, and is specifically represented as follows:

[0085] S={a1, I1, a2, I2, …, a n-1 , I n-1 , a n};

[0086] Wherein, a i is a certain letter in a-z, representing a user behavior; I i is a non-negative integer, representing the interval time between a i and a i+1 two user behaviors; with such a frequent pattern sequence based on time sequence and time slot, the user's living habits and living rules are represented.

[0087] In the above embodiment, the user behavior pattern mining based on daily routine log is carried out by using public data set, the time sequence and time interval information between behaviors are comprehensively considered, the time interval between behaviors is added as a feature element into each suffix, the suffix database containing time interval is constructed, the frequent pattern based on time sequence and time slot is mined, and the user's living habits and living rules are represented, which is more abundant and more practically meaningful than the traditional frequent pattern mining algorithm.

[0088] Corresponding to the user behavior pattern mining method based on time sequence and time slot in the above embodiment, the application further provides a user behavior pattern mining system based on time sequence and time slot, which comprises:

[0089] A user behavior sequence unit, which arranges the user behavior information in daily routine log in each day into sequence in chronological order to obtain a user behavior sequence;

[0090] A suffix database establishment unit, which constructs a suffix database containing time interval based on the user behavior sequence obtained by the user behavior sequence unit, and introduces time factor when establishing the suffix database, and adds a time interval between a prefix and a next item set into each suffix corresponding to the prefix;

[0091] A frequent pattern mining unit, which mines user behavior frequent pattern based on the suffix database established by the suffix database establishment unit, determines that the behaviors in the frequent pattern are allowed to be fuzzy in order within a time interval, and comprehensively considers the time sequence and time interval information between behaviors;

[0092] The user behavior mode representation unit represents the user's life habit and life rule based on the time sequence and time slot frequent mode obtained by the frequent mode mining unit.

[0093] For the time sequence and time slot based user behavior mode mining system of the embodiment of the present application, since it corresponds to the time sequence and time slot based user behavior mode mining method of the above embodiment, the description is relatively simple, and the relevant similar parts can be referred to the above description of the time sequence and time slot based user behavior mode mining method, which will not be described in detail here.

[0094] In the several embodiments of the present application, it should be understood that the disclosed technology can be implemented in other ways. Of course, the unit embodiment described above is only illustrative, and for example, the division of the units can be a logical function division, and in actual implementation, another division mode can be used, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, unit or module, and can be electrical or other forms.

[0095] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment.

[0096] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0097] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in other words, the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0098] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for mining user behavior patterns based on time series and time slots, characterized in that, The method includes: A suffix database containing time intervals is constructed based on user behavior sequences. When constructing the suffix database, a time factor is introduced, and a time interval between the prefix and the next item set is added to each suffix corresponding to the prefix. Based on the suffix database, the order of behaviors in frequent patterns is allowed to be ambiguous within a defined time interval. By comprehensively considering the temporality between behaviors and their time interval information, frequent patterns of user behavior are mined. Users' lifestyle habits and routines are characterized by frequent patterns based on time series and time slots; This includes comprehensively considering the temporal sequence and time interval information between behaviors to uncover frequent patterns of user behavior, including: S1. Scan the user's daily life log database to obtain the candidate itemset. Based on the set support, select the frequent itemset from the candidate itemset as the frequent k-sequence set, where k represents the sequence length in the sequence set. S2. Using each frequent 1-sequence in the frequent 1-sequence set as a prefix, establish a suffix database containing time intervals for each frequent 1-sequence. S3. Scan the constructed suffix database containing time intervals, filter frequent itemsets as suffixes of the current frequent sequence according to the set support, and perform concatenation operations with their corresponding prefixes to form a frequent k+1-sequence set. S4. Using each frequent k+1-sequence in the new frequent k+1-sequence set as a prefix, establish a suffix database containing time intervals for each frequent k+1-sequence. Repeat steps S3 and S4 above, recursively mining until all frequent sequence patterns in the daily living log database are obtained.

2. The method for mining user behavior patterns based on time series and time slots according to claim 1, characterized in that, Construct a suffix database containing time intervals, including: The start time series corresponding to the user behavior sequence is obtained from the daily routine log, as shown below: ； in, express The start time corresponding to the i-th user action, where n represents There are a total of n user actions; ; This represents a sequence of user actions, where each lowercase letter represents a user action, and the same user action is represented by the same letter. 'n' indicates... The CCP includes n user behaviors; Based on user behavior sequence and the corresponding time series Construct a sequence of triples to store user actions, corresponding times, and time intervals between adjacent actions, as shown below: ； in, n represents The CCP One triplet element; Given a sequence of triplets, obtain the suffix sequence with the i-th tuple as the prefix, as shown below: ； in, express The CCP One element; Obtain tuples based on suffix sequence The time interval suffix database with a prefix is ​​represented as follows: ； in, Indicates tuple The k-th suffix sequence with prefix , Indicates tuple The database contains a total of time intervals prefixed with suffixes. A suffix sequence.

3. The method for mining user behavior patterns based on time series and time slots according to claim 2, characterized in that, Users' lifestyle habits and routines are characterized by frequent patterns based on time series and time slots, including: Each sequence of alternating behaviors and timeslots in the frequent pattern sequence set represents a frequent behavior pattern of the user, as shown below: ； in, Take a letter from az to represent a user action; The value is a non-negative integer, representing and The time interval between these two user behaviors; a frequent pattern sequence based on time sequence and time slots, representing a user's lifestyle habits and routines.

4. A user behavior pattern mining system based on time series and time slots, characterized in that, The system includes: The user behavior sequence unit arranges the user behavior information in the daily daily log into a sequence in chronological order to obtain the user behavior sequence. The suffix database building unit constructs a suffix database containing time intervals based on the user behavior sequence obtained by the user behavior sequence unit. When building the suffix database, a time factor is introduced, and a time interval between the prefix and the next item set is added to each suffix corresponding to the prefix. The frequent pattern mining unit, based on the suffix database established by the suffix database establishment unit, allows for ambiguity in the order of behaviors in frequent patterns within a defined time interval, and comprehensively considers the temporal sequence and time interval information between behaviors to mine frequent patterns of user behavior. The user behavior pattern representation unit uses the frequent pattern representation of users' living habits and routines based on time sequence and time slot obtained by the frequent pattern mining unit. The frequent pattern mining unit comprehensively considers the temporal sequence and time interval information between behaviors to mine frequent patterns of user behavior, including: S1. Scan the user's daily life log database to obtain the candidate itemset. Based on the set support, select the frequent itemset from the candidate itemset as the frequent k-sequence set, where k represents the sequence length in the sequence set. S2. Using each frequent 1-sequence in the frequent 1-sequence set as a prefix, establish a suffix database containing time intervals for each frequent 1-sequence. S3. Scan the constructed suffix database containing time intervals, filter frequent itemsets as suffixes of the current frequent sequence according to the set support, and perform concatenation operations with their corresponding prefixes to form a frequent k+1-sequence set. S4. Using each frequent k+1-sequence in the new frequent k+1-sequence set as a prefix, establish a suffix database containing time intervals for each frequent k+1-sequence. Repeat steps S3 and S4 above, recursively mining until all frequent sequence patterns in the daily living log database are obtained.

Images