CN108134691A

CN108134691A - Model building method, Internet resources preload method, apparatus, medium and terminal

Info

Publication number: CN108134691A
Application number: CN201711364552.0A
Authority: CN
Inventors: 陈岩; 刘耀勇
Original assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd
Current assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority date: 2017-12-18
Filing date: 2017-12-18
Publication date: 2018-06-08
Anticipated expiration: 2037-12-18
Also published as: WO2019120037A1; CN108134691B

Abstract

The embodiment of the present application discloses a kind of model building method, Internet resources preload method, apparatus, medium and terminal.Wherein, the model building method includes obtaining web-based history access record and records the associated SOT state of termination with web-based history access；Record is accessed according to the web-based history and the SOT state of termination determines state transition probability matrix and observation probability matrix；Prediction model is accessed based on the state transition probability matrix and observation probability matrix structure.It can be based on current state, the observation probability of the corresponding observation of NextState is determined by accessing prediction model, so as to determine the Internet resources that subsequent time most probable accesses, it is predicted that going out target network resource before customer access network resource, the intelligent of terminal is improved.

Description

Model building method, Internet resources preload method, apparatus, medium and terminal

Technical field

The invention relates to data processing techniques more particularly to a kind of model building method, Internet resources to preload Method, apparatus, medium and terminal.

Background technology

At present, mobile terminal provides communication service, service for life, entertainment service etc. for more and more users.For example, with Family can install news category application program on mobile terminals, and Domestic News are browsed by such application program.

The application program of installation on mobile terminals provides the information of needs to the user by accessing Internet resources.But Since network loading procedure takes longer, user is caused to need to wait for network data loading, application program capacity is bad, so as to shadow Ring the usage rate of the user of the application program.By taking news category is applied as an example, word content is not only included in news interface, further includes picture Or short-sighted frequency etc..Mobile terminal is detecting that user clicks a headline, to finally showing that the headline is corresponding During news interface, by network resources address (i.e. URL, Uniform Resource Locator, the unified resource set Finger URL) few then a few tens of milliseconds (under the unobstructed environment of network) of download of network data the time it takes, at most hundreds of milliseconds Even several seconds, user experience was bad.

Invention content

The embodiment of the present application provides a kind of model building method, Internet resources preload method, apparatus, medium and terminal, The load time of network data can be shortened.

In a first aspect, the embodiment of the present application provides a kind of model building method, including：

Web-based history is obtained to access record and record the associated SOT state of termination with web-based history access；

Record is accessed according to the web-based history and the SOT state of termination determines state transition probability matrix and observation probability matrix；

Prediction model is accessed based on the state transition probability matrix and observation probability matrix structure.

Second aspect, the embodiment of the present application additionally provide a kind of Internet resources pre-add support method, including：

Current network access record is obtained according to the setting period and records associated terminal shape with current network access State；

By the access prediction model that the current network accesses record and SOT state of termination input is built in advance, wherein, it is described It is to record the corresponding SOT state of termination according to web-based history access record and web-based history access and train to access prediction model Model；

The network resource identifier to be visited of the access prediction model output is obtained, downloads the network resource identifier pair The network data answered, and it is stored in local cache.

The third aspect, the embodiment of the present application additionally provide a kind of model construction device, which includes：

Historical record acquisition module is associated with for obtaining web-based history access record and accessing record with the web-based history The SOT state of termination；

Matrix deciding module, for determining state transition probability square according to web-based history access record and the SOT state of termination Battle array and observation probability matrix；

Model construction module accesses prediction mould for being based on the state transition probability matrix and observation probability matrix structure Type.

Fourth aspect, the embodiment of the present application additionally provide a kind of Internet resources pre-load means, which includes：

Current record acquisition module, for according to setting the period obtain current network access record and with the current network Access records the associated SOT state of termination；

Mode input module, it is pre- for current network access record and the SOT state of termination to be inputted the access built in advance Model is surveyed, wherein, the access prediction model is to access record according to web-based history access record and the web-based history to correspond to The SOT state of termination training model；

Preloaded components for obtaining the network resource identifier to be visited of the access prediction model output, download institute The corresponding network data of network resource identifier is stated, and is stored in local cache.

5th aspect, the embodiment of the present application additionally provide a kind of computer readable storage medium, are stored thereon with computer Program realizes the model building method as described in above-mentioned first aspect when the program is executed by processor；Alternatively, the program is located Manage the Internet resources pre-add support method realized when device performs as described in above-mentioned second aspect.

6th aspect, the embodiment of the present application also provide a kind of terminal, including memory, processor and storage on a memory And the computer program that can be run on a processor, the processor realize such as above-mentioned first aspect when performing the computer program The model building method；Alternatively, the processor is realized when performing the computer program as described in above-mentioned second aspect Internet resources pre-add support method.

The embodiment of the present application provides a kind of model construction scheme, by obtain web-based history access record and with the history net Network access records the associated SOT state of termination；Record is accessed according to the web-based history and the SOT state of termination determines state transition probability matrix And observation probability matrix；Prediction model is accessed based on the state transition probability matrix and observation probability matrix structure；It is thus possible to Based on current state, the observation probability of the corresponding observation of NextState is determined by accessing prediction model, net is accessed in user It is predicted that going out target network resource before network resource, the intelligent of mobile terminal is improved.The embodiment of the present application also provides one kind Internet resources preload scheme, are remembered by obtaining current network according to the setting period and accessing record and accessed with the current network The associated SOT state of termination is recorded, the current network is accessed into record and the SOT state of termination inputs above-mentioned access prediction model, is treated The network resource identifier of access downloads the corresponding network data of the network resource identifier, and is stored in local cache.It can be Before obtaining network data according to a certain network resource identifier, the corresponding network number of network resource identifier that may be accessed is preloaded According to realization directly by reading network data in local cache, shortens the load time of network data, so as to avoid being added by outer net Carrying data causes the load time longer to happen.

Description of the drawings

Fig. 1 is a kind of flow chart of model building method provided by the embodiments of the present application；

Fig. 2 is the flow chart of another model building method provided by the embodiments of the present application；

Fig. 3 is a kind of flow chart of Internet resources pre-add support method provided by the embodiments of the present application；

Fig. 4 is the flow chart of another Internet resources pre-add support method provided by the embodiments of the present application；

Fig. 5 is a kind of schematic diagram of network data loading procedure provided by the embodiments of the present application；

Fig. 6 is a kind of structure diagram of model construction device provided by the embodiments of the present application；

Fig. 7 is a kind of structure diagram of Internet resources pre-load means provided by the embodiments of the present application；

Fig. 8 is a kind of structure diagram of terminal provided by the embodiments of the present application；

Fig. 9 is a kind of structure diagram of mobile terminal provided by the embodiments of the present application.

Specific embodiment

The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched The specific embodiment stated is used only for explaining the application rather than the restriction to the application.It also should be noted that in order to just Part relevant with the application rather than entire infrastructure are illustrated only in description, attached drawing.

It should be mentioned that some exemplary embodiments are described as before exemplary embodiment is discussed in greater detail The processing described as flow chart or method.Although each step is described as the processing of sequence, many of which by flow chart Step can be implemented concurrently, concomitantly or simultaneously.In addition, the sequence of each step can be rearranged.When its operation The processing can be terminated during completion, it is also possible to have the additional step being not included in attached drawing.The processing can be with Corresponding to method, function, regulation, subroutine, subprogram etc..

Fig. 1 is a kind of flow chart of model building method provided by the embodiments of the present application, and this method can be by model construction Device performs, wherein, which can generally be integrated in the terminal, such as mobile terminal by software and or hardware realization In.As shown in Figure 1, this method includes：

Step 110 obtains web-based history access record and records the associated SOT state of termination with web-based history access.

Wherein, terminal can be the mobile terminal equipped with operating system, including but not limited to smart mobile phone, tablet electricity Brain, handheld device, laptop and smartwatch etc..

Wherein, it can be that the application program in preset time section in the machine accesses network behavior that web-based history, which accesses record, Network access record, network access record include but not limited to network resources address (i.e. URL addresses), open the URL addresses Timestamp (i.e. access time) and access the URL addresses application program application identities.The SOT state of termination includes but not limited to Network environment status, charged state, remaining capacity and location information.It should be noted that the terminal shape in the embodiment of the present application Network environment status and charged state when state is application program access URL addresses etc..

When application program accesses network, mobile terminal can record network access behavior, access and record as web-based history.Together When, when recording network access behavior, the mobile terminal state that corresponding application programs access the network moment can also be obtained.Example Such as, when application program accesses network, record URL addresses, the application program for sending out network access request, the time for opening URL The network access informations such as stamp.The terminal's status informations such as the network environment of mobile terminal and charged state at this time can also be acquired.Its In, network environment can be whether to connect wireless network.Such as, if connection WIFI (Wireless Fidelity) etc..Mobile terminal can be with Web-based history access record is stored in the form of tables of data and the SOT state of termination, the tables of data are stored in mobile terminal data library.

Table 1, network access record and the tables of data that associated mobile terminal state is recorded with web-based history access

It should be noted that can be recorded a line in above table as a bar state, enumerated in above table Feature includes WIFI, application program, URL, opens the timestamp and charged state of this URL, but be not limited to features described above, can be with The features such as increase remaining capacity, location information are needed according to what realistic model was built.Opening the timestamp of this URL can be converted to Date Hour Minute Second form.For example, 14,975,906,954,69=,201,7/6,/16 13:24:55.

Step 120 determines state transition probability matrix and observation according to web-based history access record and the SOT state of termination Probability matrix.

It should be noted that mobile terminal needs to access the web-based history in above table record and the SOT state of termination carries out Pretreatment obtains state set and observation set.Wherein, state set can be according to WIFI, application program, opening URL in table 1 Timestamp and charged state determine, observation set can be determined according to URL.

It should be noted that state transition probability may be considered in state set by state q_iIt is transferred to q_jProbability, That is it is q that state transition probability, which can be understood as last moment corresponding state in the two neighboring moment,_iAnd subsequent time Corresponding state is q_jProbability.

In the embodiment of the present application, the mode for calculating the state transition probability of each state in state set is with state set By state q in conjunction_iIt is transferred to q_jNumber as molecule (it should be noted that q_iIt can be with q_jBe identical state or Different states), with state q in state set_iOccurrence number as denominator, carry out division arithmetic.It it should be noted that can To represent state set in the form of state matrix.Illustratively, equation below may be used to calculate in the state set respectively The state transition probability of a state：

Wherein, q_iRepresent the current state at current time, q_jRepresent the adjacent states of subsequent time, P (i_t+1=q_j|i_t= q_i) represent that state in the state set is transferred to the probability of adjacent states, N (i by current state_t=q_i, i_i+1=q_j) represent Current state is q in the state set_iAnd adjacent states are q_jOccurrence number, N (i_t=q_i) represent in the state set Current state is q_iNumber.

It should be noted that above-mentioned formula may be used calculates state in state set between the two neighboring moment respectively Transition probability forms state transition probability matrix according to state transition probability.For example, the state transition probability matrix can include The state that the state or the state transition probability and subsequent time of current state that subsequent time occurs occur is state matrix The state transition probability of middle residual state.If that is, represent all possible state set, i.e. Q={ q with Q₁, q₂, q₃..., q_n, wherein, n represents status number, then state transition probability matrix A can be expressed as A=[a_ij]_n×n, can be used for Record institute it is stateful between redirect (including state to itself transfer or into state set residual state transfer) probability.

It is understood that state transition probability can be deposited as shown in above-mentioned example in the form of transition transition probability matrix Can also otherwise exist and build access prediction model.For example, it can be close state and state transition probability Join the state transition probability table of storage；State transition probability is either stored in the first table, state is stored in the second table Lattice establish mapping of the first table and the second table etc..

It should be noted that observation probability may be considered the probability that preset observation is generated under a certain state, That is if with q_xRepresent a certain state, it is determined that the mode of the corresponding observation of the state can be, by state q_xLower observation The number divided by state q that value k occurs_xOccurrence number obtain state q_xThe lower observation probability for generating observation k.Illustratively, may be used Sequentially to obtain a state in state set as current state, and count the occurrence number of current state.According to described The occurrence number of the corresponding observation of current state and the current state calculates the observation of current state using equation below Probability：

Wherein, observation k ∈ { v₁,v₂,…,v_m, { v₁,v₂,…,v_mGather for observation,Represent the current shape State q_xThe lower probability for generating observation k, N (i_t=q_x, v=k) and represent the current state q_xThe lower number for generating observation k.

Step 130 accesses prediction model based on the state transition probability matrix and observation probability matrix structure.

It should be noted that the access prediction model in the embodiment of the present application can be probabilistic model.Mould is predicted in the access Type includes but not limited to hidden Markov model.Hidden Markov model describes a hiding sequence and one corresponding considerable The sequence of survey, the hiding sequence are known as status switch, can be represented by state set.The observable sequence is known as observing Sequence, can be by observing set expression.

Illustratively, if representing state set with Q, V represents observation set, then there are Q={ q₁, q₂, q₃..., q_n, V= {v₁, v₂, v₃..., v_m, wherein, n is the value range of number of states, and m is the value range of observation.Due to representing shape with A State transition probability matrix, there are A=[a_ij]_n×n, wherein, 1≤i≤n, 1≤j≤n, a_ij=P (i_t+1=q_j|i_t=q_i), a_ijTable Show by state q_iIt is transferred to q_jProbability.Observation probability matrix is the set of observation probability, that is to say, that if representing that observation is general with B Observation probability matrix can be then expressed as by rate matrixIt is thus possible to collected by state set Q, observation It closes V, state transition probability matrix A and observation probability matrix B and forms access prediction model, is i.e. the access prediction model can represent For [Q, V, A, B].

The technical solution of the present embodiment is associated with by obtaining web-based history access record and accessing record with the web-based history The SOT state of termination；Record is accessed according to the web-based history and the SOT state of termination determines state transition probability matrix and observation probability square Battle array；Prediction model is accessed based on the state transition probability matrix and observation probability matrix structure；It is thus possible to it is based on current shape State determines the observation probability of the corresponding observation of NextState by accessing prediction model, since observation represents Internet resources Address, and then, it is predicted that going out target network resource before customer access network resource, improve the intelligent of mobile terminal.

Fig. 2 is the flow chart of another model building method provided by the embodiments of the present application.As shown in Fig. 2, this method packet It includes：

Step 201 obtains web-based history access record and records the associated SOT state of termination with web-based history access.

Wherein, which accesses record and includes accessing application program, network resources address and the access time of network, And the SOT state of termination includes network environment information and charged state.

It should be noted that the acquisition of initial data (i.e. web-based history accesses record) is completed on mobile terminals. For example, SOT state of termination when obtaining the network access record of mobile terminal in real time and the network access behavior occurs, including visiting Ask WIFI when application program, network resources address (i.e. URL addresses), URL access times and the application access network of network Above-mentioned network access is recorded and the SOT state of termination is stored into the preset database of mobile terminal, deposited by state, charged state etc. Storage form can be according to shown in table 1.

Step 202, the program number according to the preset rules matching application program for accessing network and the network The corresponding network numbering of resource address.

Wherein, preset rules can be the application reference number according to preset order and be network according to preset order Resource address is numbered.Preset order can be time sequencing or other user-defined sequences.

Illustratively, with the data instance in table 1：The application program for accessing network should including QQ, wechat, today's tops etc. With for one number of each application assigned, with number alternate application program.It should be noted that for different application journeys Sequence is numbered, and identical application program, which has, to be identically numbered.For example, be QQ distribution programs number be 0, for wechat distribute journey Sequence number is 1, be today's tops distribution program number is 2 ..., and .., the maximum number of program number are depended in table 1 using journey The quantity of sequence.Application program may correspond to multiple network resources address, and different networks is distributed for different network resources address Number.Different application programs may access identical network resources address, the network numbering phase of identical network resources address Together.Each URL accessed for application program distributes a network numbering, i.e. network numbering k ∈ [0,1,2 ...], maximum network Number depends on occurring how many different URL in table 1 altogether.

Step 203 distinguishes working day and day off according to the access time, and respectively working day and day off assigns not Same date is numbered.

It should be noted that access time is the timestamp that this URL is opened in table 1, which can be converted into year Month day Hour Minute Second form, it is thus possible to be working day or rest according to the time that the access time determines to open URL addresses Day.When may be set in the time for opening URL as working day, the value of timestamp for opening this URL in table 1 is 1, if opening The time of URL is day off, then the value that the timestamp of this URL is opened in table 1 is 0.

Step 204, the period according to belonging to the access time determine time domain.

It should be noted that need respectively to be obtained the period to preset time section in consecutive days in advance, it should Period and time domain associated storage.For example, several periods were divided by 24 hours in consecutive days in advance.If for example, With 10 minutes for time interval, then at a 24 hours consecutive days there is 24*60/10=144 period, carried out for the period suitable Sequence is numbered, then access time corresponding time domain t ∈ [0,1,2,3 ... 143].Due to being time segment number, Ke Yigen The access time of the terminal access URL is determined according to the timestamp that URL is opened in a record optional in table 1, so as to according to the visit Ask the period belonging to the time and period and the time domain of advance associated storage, it may be determined that when this records corresponding Between number.

It is understood that since user can not possibly use mobile terminal in 24 hours, it can also be according to the use of user Custom, user is divided using the time interval of mobile terminal.For example, user is between 12 points of morning to 6 points of morning In sleep state, mobile terminal will not be used, then can be drawn to excluding the time interval except this time of having a rest section Point, obtain the period.

Step 205 judges whether terminal accesses wireless network according to the network environment information, is determined according to judging result Network-like state value.

Wherein, the information such as network, packet are accessed whether network environment information includes accessing terminal to network and using which kind of approach It includes but is not limited to whether access WIFI.

If can prespecified mobile terminal access wireless network, it is 1 to set network-like state value, if mobile terminal does not connect Enter wireless network, then it is 0 to set network-like state value.For example, if the WIFI of mobile terminal is opened and access wireless network, set Network-like state value is 1, if the WIFI of mobile terminal is closed, it is 0 to set network-like state value.

Step 206 judges whether terminal is charging according to charged state, and state-of-charge value is determined according to judging result.

If can prespecified mobile terminal be in charged state, state-of-charge value is set for 1, if mobile terminal is not located In charged state, then state-of-charge value is 0.Whether charged according to mobile terminal, determining charging when application program accesses URL State value.

Following table can be obtained according to the above-mentioned pretreatment for accessing web-based history record and the SOT state of termination.

Table 2, sample set table

WIFI	Application program	Whether working day	Period	Charged state	URL (network address)
						1	0	1	81	1	0
0	1	1	82	0	1
						…	…	…	…	…	…
…	…	…	…	…	…

It should be noted that a state is represented per a line in table 2.

Step 207 is numbered, the generation of date codes, time domain, network-like state value and state-of-charge value according to described program State set, and observation set is formed according to the network numbering.

It should be noted that numbered by described program, date codes, time domain, network-like state value and state-of-charge value State set is formed, observation set is formed by the network numbering.

Illustratively, by taking the sample set in table 2 as an example, state set Q is formed by preceding 5 row character pair value, by last One row URL forms observation set V.Due in model construction, obtain the web-based history of at least two weeks and access and record and corresponding The SOT state of termination so the randomness of the record in the corresponding state matrix of state set is very big, is not necessarily sequentially in time Arrangement.Optionally, state set Q={ q₁=[0,0,0,0,0], q₂=[0,0,0,0,1] ..., q_i=[1,0,1,81,1], q_j=[0,1,1,82,0] ..., q_n, observation set V={ 0,1,2 ..., v_m}.If state set and sight are represented in the matrix form Surveying set can be as follows：

And

Wherein, the record number of Q can be expressed as accessing the number of the application program of network in 2*2*144*2* terminals, V's Size is to access the number of URL.

Step 208, the state transition probability for calculating each state in the state set, and shifted generally according to the state Rate generates state transition probability matrix.

It should be noted that state transition probability in the embodiment of the present application calculated is that current state and next line can The transition probability of state that can occur, i.e., current state occurs in pairs with NextState in above-mentioned state set, and next shape State is located at the next line that current state is expert at.For example, under certain conditions, current state is q_i=[1,0,1,81,1], and NextState is q_j=[0,1,1,82,0], and under other states, current state is q_i=[1,0,1,81,1], but under One state q_j=[0,0,0,0,1] etc., can be by the above-mentioned two kinds of q enumerated_jIt is considered adjacent next shape of current state State.Since each user has the network access custom of oneself, if a certain period user at certain day has accessed some URL, then Within the sampling period of at least two weeks, the probability that user accesses similary URL in same time period can be higher, therefore, in shape There is the state repeated in state matrix.Illustratively, to calculate by current state q_iJump to NextState q_jState Transition probability, then it is q that can count lastrow in above-mentioned state matrix Q_iAnd next line is q_jState to quantity, in addition, also uniting It counts in above-mentioned state matrix Q and q occurs_iThe quantity of state.So as to be q by lastrow in state matrix Q_iAnd next line is q_jShape State is to quantity divided by q_iThe result of the quantity of state is used as by current state q_iJump to NextState q_jState transition probability. It should be noted that q_jIncluding with removing q in state matrix_iResidual state, further include state q_iItself.For state matrix In each state computation its state transition probability, obtain the state transition probability matrix of n*n, can be above-mentioned [a_ij]_n×n。

Step 209, the observation probability that the corresponding observation of each state is calculated according to the state set and observation set, And observation probability matrix is generated according to the observation probability.

In the embodiment of the present application, a line state in sequence acquisition state matrix obtains and current as current state The corresponding observation of state counts the occurrence number N (i of identical observation_t=q_x, v=k), wherein, x ∈ [1,2,3 ..., N], k ∈ [1,2,3 ..., v_m].Then, by N (i_t=q_x, v=k) divided by state matrix in current state occurrence number fortune Result is calculated as the observation probability by generating observation k under current state

According to aforesaid way, using each state in state matrix as current state, determine to produce under current state respectively The probability of observation in raw observation set forms the observation probability matrix B of n*m.

Step 210, structure access prediction model.

In the embodiment of the present application, it is hidden Markov model to access prediction model.Gathered according to state set, observation, State transition probability matrix A and observation probability matrix B may be constructed hidden Markov model λ, and accessing prediction model can represent For λ=[Q, V, A, B], which is built in terminal, is desirable for prejudging the corresponding resources of which URL in advance It downloads in advance.

It should be noted that since user uses the application program on mobile terminal to access the randomness of Internet resources very Greatly, the sample set of network resources address that may be accessed for predicting mobile terminal future be not it is unalterable, can be with The update condition of model is set according to actual needs, so as to when meeting update condition, be carried out more to the access prediction model Newly.Illustratively, renewal time is set, when detecting that system time meets the renewal time, is obtained in preset time section Web-based history access behavior and status information of mobile terminal, be updated to accessing prediction model.Can be specifically to set more The new period when reaching the update cycle, determines to meet model modification condition.For example, one week or one day can be set as update Period.If updating a model weekly, detecting that by the way of incremental update, acquisition is nearest when meeting update interval The network behavior record and the SOT state of termination of mobile terminal in one week.It is understood that update date can also be set, such as often Monday or week week about are first-class.Sample Refreshment operation can also be triggered by user, for example, the update of user's input model refers to Show, model modification operation etc. is performed based on the update instruction.

The technical solution of the present embodiment accesses behavior by the web-based history of learning terminal and accesses behavior with web-based history The associated SOT state of termination, structure access prediction model, improve the adaptive ability of model, can have to different user good Applicability.

Fig. 3 is a kind of flow chart of Internet resources pre-add support method provided by the embodiments of the present application, and this method can be by net Network resource pre-load means perform, wherein, which can generally be integrated in the terminal, example by software and or hardware realization In mobile terminal.As shown in figure 3, this method includes：

Step 310 obtains current network access record according to the setting period and is associated with current network access record The SOT state of termination.

It needs, current network accesses record and includes current time with accessing the application program of network, Internet resources Location (i.e. URL addresses) and the timestamp for opening the URL.The associated SOT state of termination, which is recorded, with current network access includes current net Network environmental information and present charge state, including but unlimited monitor terminal currently whether open access WIFI and monitor terminal know Not in charging etc..

It should be noted that the setting period can be system default, can also be obtain user setting for calling Access the time interval that prediction model predicts NextState according to current state.For example, can set one was called every 10 minutes Secondary access prediction model with the network resources address that may be accessed according to current state prediction NextState, and is downloaded pre- in advance The associated network data of network resources address measured.

It is understood that after getting current network and accessing record and its associated SOT state of termination, need to current Network access record and the SOT state of termination are pre-processed, and (mode of the pretreatment is identical with the record of above-described embodiment, herein not Repeat again), obtain the current state of the input data format match with accessing prediction model.

Step 320, the access prediction model for building current network access record and SOT state of termination input in advance.

Wherein, the access prediction model be according to web-based history access record and the web-based history access record it is corresponding The model of SOT state of termination training.It is understood that access prediction model herein can be recorded by above-described embodiment The model of scheme constructs.

In the present embodiment, due to being pre-processed to obtain current shape to current network access record and the SOT state of termination Current state is directly exported the access prediction model by state, is likely to occur with prejudging subsequent time by the access prediction model NextState, and then, calculate NextState corresponding each observation (i.e. network resource identifier, also, Internet resources respectively Mark can include network resources address) observation probability, the observation probability may be considered access prediction model output.Show Example property, by current state q_iInput accesses prediction model, and current state q is determined according to state transition probability matrix_iCorresponding institute There is the state transition probability of next state, so as to which a state of state transition probability maximum most may be used as subsequent time The NextState q that can occur_j.Then, further according to q_jAnd observation probability matrix B determines q_jIt is corresponding to refer to observation probability.To the ginseng It examines observation probability and carries out descending arrangement, it is general with reference to the target observation that the preceding setting quantity that sorts is filtered out in observation probability by this Rate, the network resource identifier that the corresponding network resource identifier of target observation probability is accessed as NextState most probable, goes forward side by side Row output.

Step 330 obtains the network resource identifier to be visited for accessing prediction model output, downloads the network money Source identifies corresponding network data, and is stored in local cache.

It should be noted that the network resource identifier can be network resources address.For example, the above-mentioned quantity that sets is 100, Then obtain corresponding 100 network resources address of NextState of access prediction model output, this 100 network resources address It is regarded as needing to preload.

It is understood that after prediction obtains network resources address, if the corresponding network number of the network resources address According to being loaded, then without repeating to load the network data, avoid excessively occupying caching.Illustratively, it is obtained in prediction above-mentioned After 100 network resources address, judge that prediction obtains whether the corresponding network data of the network resources address has been preloaded； If so, abandoning performing the down operation for the network data, otherwise, the corresponding network number of the network resource identifier is downloaded According to, and it is stored in local cache.

It should be noted that the mode for downloading the corresponding network data of the network resource identifier can be by internet with The corresponding server of network resources address establishes communication connection, and the corresponding network number of the network resources address is downloaded by the server According to.It is understood that the mode for downloading the corresponding network data of the network resource identifier is not limited to the side that above-mentioned example is enumerated Formula.

The technical solution of the present embodiment, by according to setting the period obtain current network access record and with the current net Network access records the associated SOT state of termination, and the current network is accessed record and the SOT state of termination inputs above-mentioned access and predicts mould Type obtains network resource identifier to be visited, downloads the corresponding network data of the network resource identifier, and is stored in local slow It deposits, the network resource identifier pair that may be accessed can be preloaded before network data is obtained according to a certain network resource identifier The network data answered is realized directly by reading network data in local cache, shortens the load time of network data, so as to keep away Exempt from the load time to be caused longer to happen by outer net loading data.

It should be noted that the network data of above-mentioned preloading is buffered in local cache, can be obtained in application program During to network access request, directly by downloading the corresponding network data of the network access request in local cache.Wherein, using journey Sequence includes the application program with network access functions.For example, news category application program, instant messaging class application program or more matchmakers Body class application program etc..Obtain the network access request of application program.Wherein, network access request is appreciated that as by applying journey Sequence is to the message of the corresponding server acquisition request network data of specified network resources address (i.e. URL).The network access request It can be generated by application program, and report to grid function module.The request is got in grid function module to obtain When fetching determines the message of the network data of URL, system has learnt that application program will obtain Internet resources.It should be noted that it is System network function module can be understood as system and realize code and protocol stack of network access functions etc..

Then, the local cache is inquired according to the network access request, determines that the network access request is corresponding The preload condition of network data.Illustratively, network access request includes network resources address (URL addresses), should obtaining During with the network access request of program, network resources address is extracted.Local cache is inquired according to the network resources address, judgement is It is no that there are the corresponding network datas of the network resources address.If there is no the corresponding network data of the network resources address, really The fixed corresponding network data of the network access request does not preload, performs network access flow, the network access is downloaded by outer net Ask corresponding network data.If there are the corresponding network datas of the network resources address, it is determined that the network access request pair The network data answered has preloaded, and by reading the network data in local cache, is directly returned to send out network access request Application program avoids mobile terminal from, by server download of network data, shortening the time of network data loading by internet.

Fig. 4 is the flow chart of another Internet resources pre-add support method provided by the embodiments of the present application.It as shown in figure 4, should Method includes：

Step 401 obtains current network access record according to the setting period and is associated with current network access record The SOT state of termination.

Wherein, the setting period can be system default or user is set according to actual needs, for triggering root The operation of network resource identifier that may be accessed according to current time network access record and SOT state of termination prediction subsequent time.

Step 402 pre-processes current network access record and the SOT state of termination, obtains current state.

Step 403, the access prediction model for building current state input in advance, to predict mould by described access The corresponding NextState of type prediction subsequent time, and determine that the observation of the corresponding network resources address of the NextState is general respectively Rate.

Step 404 obtains the corresponding network resources address to be visited of NextState for accessing prediction model output.

Step 405 judges whether terminal accesses wireless network, if so, performing step 406, otherwise performs step 407.

Wireless network in the embodiment of the present application includes but not limited to WIFI, that is to say, that is getting access prediction model After the network resources address of output, the networking state of terminal can be obtained, whether judgement terminal connects WIFI.

Step 406 downloads the corresponding network data of the network resource identifier, and is stored in local cache.

It is understood that in certain embodiments, after the completion of this step performs, it can redirect and perform step 413.At certain In a little embodiments, after the completion of this step performs, step 410 can be jumped to, performs step 410 to be identified in down loading network resource After corresponding network data, the first network resource address for being connected to the network resource identifier and including is obtained by web crawlers The second network resources address, load the corresponding network data of the second network resources address, and be stored in local cache.

The status information of step 407, the cellular mobile network data of the acquisition terminal.

Wherein, in the present embodiment status information include but not limited to cellular mobile network data switch on off state, The daily flow upper limit, the same day have used flow and residual flow state.

Step 408 judges whether the status information meets preset condition, if so, performing step 406, otherwise performs Step 409.

It should be noted that preset condition, which can be cellular mobile network data switch, is in opening and residual flow More than preset flow thresholding.Optionally, preset condition can also be that cellular mobile network data switch is in opening and works as It is preceding that flow has been used to be not up to preset daily flow upper limit etc..

Step 409 determines mesh according to the data type and/or data volume of the corresponding network data of the network resource identifier Network data is marked, the target network data are loaded.

It should be noted that when the status information of cellular mobile network data is to be unsatisfactory for preset condition, obtain by visiting Ask the data type of the corresponding network data of network resource identifier of prediction model output, wherein, which is included but not It is limited to word, picture, animation, video and audio etc..Optionally, the corresponding network data of the network resource identifier can also be obtained Data volume (the namely size of network data).For example, the data volume of network data D1 is 5MB, the data of network data D2 Amount is 50MB etc..Target network data can refer to the network data for meeting default screening conditions, wherein, screening conditions can be Data type can be used as target network data for the network data of word and picture.Optionally, screening conditions can also be several Target network data can be used as according to network data of the amount less than preset data amount threshold value.

The data type of the corresponding network data of the network resource identifier of acquisition is matched with default screening conditions, if Successful match, it is determined that the network data of successful match is target network data, and is loaded to the target network data.It can Choosing, the data volume of the corresponding network data of the network resource identifier of acquisition is matched with default screening conditions, if matching Success, it is determined that the network data of successful match is target network data, and is loaded to the target network data.It can manage Solution, above-mentioned data type and data volume can combine for judging target network data, can also be used alone.

Step 410 is linked to the first network resource address that the network resource identifier includes by web crawlers acquisition The second network resources address, load the corresponding network data of second network resources address, and be stored in local cache.

It it should be noted that may comprising other Internet resources in the corresponding network data of certain network resources address The link of location can be obtained these additional network resources addresses by web crawlers, and carry out data preloading.For example, under When carrying target network data, find there is the second network resources address (second network for being linked to first network resource address Resource address can be not belonging to access the set of the network resource identifier of prediction model output), second net can be downloaded together The corresponding network data of network resource address.It should be noted that equally can basis when downloading second network resources address Default screening conditions described in step 409 determine the network data that need to actually download.

It is understood that this step it is not necessary to perform the step of, in certain embodiments in execution of step 409 After can not also perform this step.

Step 411 obtains the memory space that network data occupies in the local cache.

It is understood that it due to preload web data and is stored in local cache, with the network number of preloading According to the increase of amount, occupancy local cache is also more, may influence the operational efficiency of terminal.In order to avoid local cache is too many The partial data in setting strategy removing local cache may be used, to reduce the network stored in local cache in network data Data volume.It should be noted that triggering is performed there are many kinds of the modes of data dump event, the embodiment of the present application is not made specifically It limits.For example, the size that the data volume of network data in local cache can be obtained according to the setting period (is namely occupied and is stored The size in space).For another example, the time for removing network data caching can be preset, local delay is carried out at 0 point~1 point such as morning Removing of network data deposited etc..

Step 412, when the memory space is more than given threshold, according to the cache-time of the network data and/or Frequency of use determines data to be cleaned, and removes the data to be cleaned.

Wherein, given threshold can be system default, can also according to actual needs be set by user.It is exemplary , if total buffer data size is more than 100MB, remove setting cache file.Wherein, setting cache file can be according to caching The cache-time of file determines, for example, data cached more than 1 hour of cache-time is considered setting cache file. In addition, setting cache file can also be determined according to frequency of use, for example, frequency of use can less than the data cached of 5 times/hour Be considered as setting cache file.It should be noted that above-mentioned cache-time and frequency of use can be integrated for true Surely cache file is set, can also be used alone.

It should be noted that the execution sequence sequence that it is not limited to the above example records of step 411 and step 412 is (i.e. After step 410), as long as detecting that data dump event is triggered performs step 411 and step 412.

Step 413, the network access request for obtaining application program.

It should be noted that the execution order of this step does not limit, the net inputted in a certain application program by user The operation triggering of network access request.That is monitoring user operation, when user inputs network access request operation, terminal obtains application The network access request of program.

Step 414 inquires the local cache according to the network access request, determines that the network access request corresponds to Network data preload condition.

Wherein, preload condition includes being loaded onto local cache and not loaded.It is included according to network access request Network resources address inquires local cache, judges whether the corresponding network data of the network resources address has loaded, if so, holding Otherwise row step 415 performs step 416.

Step 415 reads the network data, and is back to the application program.

Step 416 downloads the network data by outer net.

Technical solution provided in this embodiment, by before down loading network resource identifies corresponding network data to terminal Network environment prejudged, if under wireless network environment (such as under WIFI environment), can direct preload web number According in the network access request for detecting application program input by user, directly to read network data by local cache, carry Rise data loading.If under cellular mobile network data environment, according to setting policy download all or part network number According to satisfaction promotes data loading and excessive flow is not caused to expend simultaneously.In addition, it is exported in download access prediction model The corresponding network data of network resource identifier except, also load the corresponding network number of other links that the network data includes According to abundant preloading data avoids data from omitting.Meanwhile it by formulating appropriate data cached purge mechanism, avoids local slow It deposits too many network data and influences the execution efficiency of terminal.

Fig. 5 is a kind of schematic diagram of network data loading procedure provided by the embodiments of the present application.As shown in figure 5, terminal Network subsystem (being referred to as grid function module) according to setting period current network access record and with it is current Network access records the associated SOT state of termination, and utilizes and access prediction model (can be hidden Markov model) based on current net Network accesses record and SOT state of termination prediction next network resources address that most possibly accesses, using the network resources address as Decision is exported.That is the output for accessing prediction model is the network of setting quantity that the following most probable of terminal accesses Resource address.The corresponding network data of the network resources address is downloaded by server in outer net according to the network resources address, and Network data is stored in local cache, the corresponding content deposit local caches of the URL that will be preloaded.If it currently should detect The network access request sent out with program then extracts the network resources address that the network access request includes, and judges that the network provides Whether the corresponding content of source address has been preloaded into local.If having preloaded, by with reading the Internet resources in local cache The corresponding cache file in location, and the application program is directly returned to, without by server download of network data in outer net.But If not preloading the corresponding network data of the network resources address, network access request continues, and being downloaded by external network server should The corresponding network data of network resources address.

The technical solution of the present embodiment is accessed prediction model by being called according to the setting period, is accessed based on current network It records and records the associated SOT state of termination with current network access, with predicting the Internet resources that following most probable accesses Location, and the corresponding network data of the network resources address is downloaded in advance, it can be in the network access request for detecting application program When, judge the network access request network resources address whether in local cache there are corresponding network data, if so, directly It connects by reading the corresponding network data of network access request in local cache, when shortening loading of the application program for Internet resources Between, the performance of application program is improved, avoids the problem that causing the load time longer by outer net loading data.

Fig. 6 is a kind of structure diagram of model construction device provided by the embodiments of the present application.The device can pass through software And/or hardware realization, it can be integrated in terminal, for performing model building method provided by the embodiments of the present application.Such as Fig. 6 institutes Show, which includes：

Historical record acquisition module 610 records for obtaining web-based history access record and being accessed with the web-based history The associated SOT state of termination；

Matrix deciding module 620, for determining that state transfer is general according to web-based history access record and the SOT state of termination Rate matrix and observation probability matrix；

Model construction module 630 accesses in advance for being based on the state transition probability matrix and observation probability matrix structure Survey model.

The technical solution of the present embodiment provides a kind of model construction device, by obtain web-based history access record and with this Web-based history access records the associated SOT state of termination；Record is accessed according to the web-based history and the SOT state of termination determines that state transfer is general Rate matrix and observation probability matrix；Prediction model is accessed based on the state transition probability matrix and observation probability matrix structure；From And current state can be based on, the observation probability of the corresponding observation of NextState is determined by accessing prediction model, due to seeing Measured value represents network resources address, and then, it is predicted that going out target network resource before customer access network resource, improve shifting Move the intelligent of terminal.

Optionally, matrix deciding module 620 includes：

Submodule is pre-processed, for being pre-processed to web-based history access record and the SOT state of termination, obtains state Set and observation set；

First computational submodule, for calculating the state transition probability of each state in the state set, and according to institute State state transition probability generation state transition probability matrix；

Second computational submodule, for calculating the corresponding observation of each state according to the state set and observation set Observation probability, and according to the observation probability generate observation probability matrix.

Optionally, the state transition probability that equation below calculates each state in the state set may be used：

Optionally, the second computational submodule is specifically used for：

Sequence obtains a state in the state set as current state, and count current state goes out occurrence Number；

According to the corresponding observation of the current state and the occurrence number of the current state, using equation below meter Calculate the observation probability of current state：

Optionally, the web-based history accesses record and includes accessing application program, network resources address and the access of network Time and, the SOT state of termination include network environment information and charged state.

And pretreatment submodule is specifically used for：

According to the program number of the preset rules matching application program for accessing network and the network resources address Corresponding network numbering；

Working day and day off are distinguished according to the access time, respectively working day and day off assigns not same date and compile Number；

Period according to belonging to the access time determines time domain, wherein, to preset time section in consecutive days Respectively obtained the period, the period and time domain associated storage；

Judge whether terminal accesses wireless network according to the network environment information, network state is determined according to judging result Value；

Judge whether terminal is charging according to charged state, state-of-charge value is determined according to judging result；

Numbered according to described program, date codes, time domain, network-like state value and state-of-charge value generating states set It closes, and observation set is formed according to the network numbering.

Fig. 7 is a kind of structure diagram of Internet resources pre-load means provided by the embodiments of the present application.The device can lead to Software and or hardware realization is crossed, can be integrated in terminal, is preloaded for performing Internet resources provided by the embodiments of the present application Method.As shown in fig. 7, the device includes：

Current record acquisition module 710, for according to the setting period obtain current network access record and with it is described current Network access records the associated SOT state of termination；

Mode input module 720, for the visit for building current network access record and SOT state of termination input in advance Ask prediction model, wherein, the access prediction model is to access record and web-based history access record according to web-based history The model of corresponding SOT state of termination training；

Preloaded components 730 for obtaining the network resource identifier to be visited of the access prediction model output, are downloaded The corresponding network data of the network resource identifier, and it is stored in local cache.

The technical solution of the present embodiment provides a kind of Internet resources pre-load means, can be according to a certain Internet resources mark Know before obtaining network data, preload the corresponding network data of network resource identifier that may be accessed, realize directly by local Network data is read in caching, shortens the load time of network data, so as to avoid leading to the load time by outer net loading data Longer happens.

Optionally, mode input module 720 is specifically used for：

Record is accessed to the current network and the SOT state of termination pre-processes, obtains current state；

The access prediction model that current state input is built in advance, with by under the access prediction model prediction One moment corresponding NextState, and the observation probability of the corresponding network resources address of the NextState is determined respectively.

Optionally, it further includes：

Download policy selecting module, for before the corresponding network data of the network resource identifier is downloaded, judging end Whether end accesses wireless network；

If so, perform the operation for downloading the corresponding network data of the network resource identifier；

Otherwise, the status information of the cellular mobile network data of the terminal is obtained；

When the status information meets preset condition, perform and download the corresponding network data of the network resource identifier Operation；

And when the status information is unsatisfactory for preset condition, preloaded components 730 are specifically used for：

Target network number is determined according to the data type of the corresponding network data of the network resource identifier and/or data volume According to being loaded to the target network data.

Optionally, it further includes：

Address acquisition module, for after the corresponding network data of the network resource identifier is downloaded, being climbed by network Worm, which obtains, is linked to the second network resources address of the first network resource address that the network resource identifier includes, described in loading The corresponding network data of second network resources address, and it is stored in local cache.

Optionally, it further includes：

Caching removes module, for obtaining the memory space that network data occupies in the local cache；In the storage When space is more than given threshold, data to be cleaned are determined, and clear according to the cache-time of the network data and/or frequency of use Except the data to be cleaned.

Optionally, it further includes：

Acquisition request module, for obtaining the network access request of application program；

Caching query module for inquiring the local cache according to the network access request, determines that the network is visited Ask the preload condition for asking corresponding network data；

Data read module, if having been preloaded for the corresponding network data of the network access request, described in reading Network data, and it is back to the application program；

Data download module, if not preloaded for the corresponding network data of the network access request, by outer off the net Carry the network data.

The embodiment of the present application also provides a kind of storage medium for including computer executable instructions, and the computer can perform When being performed by computer processor for performing a kind of model building method, this method includes for instruction：

In addition, the embodiment of the present application also provides another storage medium for including computer executable instructions, the calculating When being performed by computer processor for performing a kind of Internet resources pre-add support method, this method includes machine executable instruction：

Storage medium --- any various types of memory devices or storage device.Term " storage medium " is intended to wrap It includes：Install medium, such as CD-ROM, floppy disk or magnetic tape equipment；Computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, blue Bath (Rambus) RAM etc.；Nonvolatile memory, such as flash memory, magnetic medium (such as hard disk or optical storage)；Memory component of register or other similar types etc..Storage medium can further include other The memory or combination of type.In addition, storage medium can be located at program in the first computer system being wherein performed, Or can be located in different second computer systems, second computer system is connected to the by network (such as internet) One computer system.Second computer system can provide program instruction and be used to perform to the first computer." storage is situated between term Matter " can include may reside in different location two of (such as in different computer systems by network connection) or More storage mediums.Storage medium can store the program instruction that can be performed by one or more processors and (such as implement For computer program).

Certainly, a kind of storage medium for including computer executable instructions that the embodiment of the present application is provided, computer The mould that the application any embodiment is provided can also be performed in the operation for the model construction that executable instruction is not limited to the described above Relevant operation in type construction method.

It should be noted that a kind of storage medium for including computer executable instructions that the embodiment of the present application is provided, The operation that the Internet resources that its computer executable instructions is not limited to the described above preload, it is arbitrarily real to can also be performed the application Apply the relevant operation in the Internet resources pre-add support method that example is provided.

The embodiment of the present application provides a kind of terminal, and model construction dress provided by the embodiments of the present application can be integrated in the terminal It puts.The embodiment of the present application provides another terminal, and Internet resources pre-add provided by the embodiments of the present application can be integrated in the terminal It carries and puts.It should be noted that model construction can be carried out by the first terminal, and model transplantations are pre- to Internet resources are performed Second of terminal of operation is loaded, model construction and Internet resources preload operation can also be performed in same terminal.Wherein, Terminal includes smart mobile phone, tablet computer, handheld device, laptop and smartwatch etc..Fig. 8 is the embodiment of the present application A kind of structure diagram of the terminal provided.As shown in figure 8, the terminal can include：Memory 810 and processor 820.It is described to deposit Reservoir 810, for storing computer program, web-based history access record, the SOT state of termination and accessing prediction model；The processor 820 read and perform the computer program stored in the memory 810.The processor 820 is performing the computer journey Following steps are realized during sequence：Web-based history is obtained to access record and record the associated SOT state of termination with web-based history access； Record is accessed according to the web-based history and the SOT state of termination determines state transition probability matrix and observation probability matrix；Based on described State transition probability matrix and observation probability matrix structure access prediction model.

It should be noted that the memory 810 of terminal can be also used for storing another computer program and preloading Network data.Processor 820 can be also used for reading and perform the computer program that is stored in the memory 810.The place Reason device 820 realizes following steps when performing the computer program：According to setting the period obtain current network access record and The associated SOT state of termination is recorded with current network access；The current network is accessed into record and SOT state of termination input is advance The access prediction model of structure, wherein, the access prediction model is to access record and the web-based history according to web-based history Access the model for recording corresponding SOT state of termination training；Obtain the Internet resources mark to be visited of the access prediction model output Know, download the corresponding network data of the network resource identifier, and be stored in local cache.

The memory and processor enumerated in above-mentioned example are the part component of terminal, and the terminal can also include Other components.By taking mobile terminal as an example, illustrate the possible structure of above-mentioned terminal.Fig. 9 is one kind provided by the embodiments of the present application The structure diagram of mobile terminal.As shown in figure 9, the mobile terminal can include：Memory 901, central processing unit (Central Processing Unit, CPU) 902 (also known as processor, hereinafter referred to as CPU), Peripheral Interface 903, RF (Radio Frequency, radio frequency) circuit 905, voicefrequency circuit 906, loud speaker 911, power management chip 908, input/output (I/O) son System 909, other input/control devicess 910 and outside port 904, these components by one or more communication bus or Signal wire 907 communicates.

It should be understood that diagram mobile terminal 900 is only an example of mobile terminal, and mobile terminal 900 Can have than more or less components shown in figure, two or more components can be combined or can be with It is configured with different components.Various parts shown in figure can be including one or more signal processings and/or special Hardware, software including integrated circuit are realized in the combination of hardware and software.

By taking the mobile terminal is mobile phone as an example, mobile terminal is described in detail.

Memory 901, the memory 901 can be by access such as CPU902, Peripheral Interfaces 903, and the memory 901 can To include high-speed random access memory, nonvolatile memory can also be included, such as one or more disk memory, Flush memory device or other volatile solid-state parts.Computer program is stored in memory 901, history can also be stored Network access record, the SOT state of termination, access prediction model and network data of preloading etc..

The peripheral hardware that outputs and inputs of equipment can be connected to CPU902 and deposited by Peripheral Interface 903, the Peripheral Interface 903 Reservoir 901.

I/O subsystems 909, the I/O subsystems 909 can be by the input/output peripherals in equipment, such as touch screen 912 With other input/control devicess 910, it is connected to Peripheral Interface 903.I/O subsystems 909 can include 9091 He of display controller For controlling one or more input controllers 9092 of other input/control devicess 910.Wherein, one or more input controls Device 9092 processed receives electric signal from other input/control devicess 910 or sends electric signal to other input/control devicess 910, Other input/control devicess 910 can include physical button (pressing button, rocker buttons etc.), dial, slide switch, behaviour Vertical pole clicks idler wheel.What deserves to be explained is input controller 9092 can with it is following any one connect：Keyboard, infrared port, The indicating equipment of USB interface and such as mouse.

Touch screen 912, the touch screen 912 are the input interface and output interface between user terminal and user, can User is shown to depending on output, visual output can include figure, text, icon, video etc..

Display controller 9091 in I/O subsystems 909 receives electric signal from touch screen 912 or is sent out to touch screen 912 Electric signals.Touch screen 912 detects the contact on touch screen, and the contact detected is converted to and shown by display controller 9091 The interaction of user interface object on touch screen 912, that is, realize human-computer interaction, the user interface being shown on touch screen 912 Icon that object can be the icon of running game, be networked to corresponding network etc..What deserves to be explained is equipment can also include light Mouse, light mouse are the extensions for not showing the touch sensitive surface visually exported or the touch sensitive surface formed by touch screen.

RF circuits 905 are mainly used for establishing the communication of mobile phone and wireless network (i.e. network side), realize mobile phone and wireless network The data receiver of network and transmission.Such as transmitting-receiving short message, Email etc..Specifically, RF circuits 905 receive and send RF letters Number, RF signals are also referred to as electromagnetic signal, and RF circuits 905 convert electrical signals to electromagnetic signal or electromagnetic signal is converted to telecommunications Number, and communicated by the electromagnetic signal with communication network and other equipment.RF circuits 905 can include performing The known circuit of these functions includes but not limited to antenna system, RF transceivers, one or more amplifiers, tuner, one A or multiple oscillators, digital signal processor, CODEC (COder-DECoder, coder) chipset, user identifier mould Block (Subscriber Identity Module, SIM) etc..

Voicefrequency circuit 906 is mainly used for receiving audio data from Peripheral Interface 903, which is converted to telecommunications Number, and the electric signal is sent to loud speaker 911.

Loud speaker 911 for the voice signal for receiving mobile phone from wireless network by RF circuits 905, is reduced to sound And play the sound to user.

Power management chip 908, the hardware for being connected by CPU902, I/O subsystem and Peripheral Interface are powered And power management.

Terminal provided by the embodiments of the present application can be based on current state, NextState is determined by accessing prediction model The observation probability of corresponding observation it is predicted that going out target network resource before customer access network resource, improves movement Terminal it is intelligent.

Terminal provided by the embodiments of the present application, can be before network data be obtained, in advance according to a certain network resource identifier The corresponding network data of network resource identifier that may be accessed is loaded, realizes directly by reading network data in local cache, contracts The load time of short network data, so as to avoid the load time being caused longer to happen by outer net loading data.

Model construction device, storage medium and the terminal provided in above-described embodiment can perform the embodiment of the present application and be provided Model building method, have and perform the corresponding function module of this method and advantageous effect.It is not detailed in the above-described embodiments to retouch The technical detail stated, reference can be made to the model building method that the embodiment of the present application is provided.

Internet resources pre-load means, storage medium and the terminal provided in above-described embodiment can perform the embodiment of the present application The Internet resources pre-add support method provided has and performs the corresponding function module of this method and advantageous effect.Not in above-mentioned reality The technical detail of detailed description in example is applied, reference can be made to the Internet resources pre-add support method that the embodiment of the present application is provided.

Note that it above are only the preferred embodiment of the application and institute's application technology principle.It will be appreciated by those skilled in the art that The application is not limited to specific embodiment described here, can carry out for a person skilled in the art it is various it is apparent variation, The protection domain readjusted and substituted without departing from the application.Therefore, although being carried out by above example to the application It is described in further detail, but the application is not limited only to above example, in the case where not departing from the application design, also It can include other more equivalent embodiments, and scope of the present application is determined by scope of the appended claims.

Claims

1. a kind of model building method, which is characterized in that including：

It is 2. according to the method described in claim 1, it is characterized in that, true according to web-based history access record and the SOT state of termination Determine state transition probability matrix and observation probability matrix, including：

Record is accessed to the web-based history and the SOT state of termination pre-processes, obtains state set and observation set；

The state transition probability of each state in the state set is calculated, and state is generated according to the state transition probability and is turned Move probability matrix；

The observation probability of the corresponding observation of each state is calculated according to the state set and observation set, and according to the sight Survey probability generation observation probability matrix.

3. according to the method described in claim 2, it is characterized in that, calculate the state transfer of each state in the state set Probability, including：

The state transition probability of each state in the state set is calculated using equation below：

Wherein, q_iRepresent the current state at current time, q_jRepresent the adjacent states of subsequent time, P (i_t+1=q_j|i_t=q_i) table Show that state in the state set is transferred to the probability of adjacent states, N (i by current state_t=q_i, i_i+1=q_j) represent the shape Current state is q in state set_iAnd adjacent states are q_jOccurrence number, N (i_t=q_i) represent current shape in the state set State is q_iNumber.

4. according to the method described in claim 3, it is characterized in that, each shape is calculated according to the state set and observation set The observation probability of the corresponding observation of state, including：

A state in the state set is sequentially obtained as current state, and counts the occurrence number of current state；

According to the corresponding observation of the current state and the occurrence number of the current state, calculated and worked as using equation below The observation probability of preceding state：

Wherein, observation k ∈ { v₁,v₂,…,v_m, { v₁,v₂,…,v_mGather for observation,Represent the current state q_xUnder Generate the probability of observation k, N (i_t=q_x, v=k) and represent the current state q_xThe lower number for generating observation k.

5. method according to any one of claim 2 to 4, which is characterized in that the web-based history accesses record and includes Access network application program, network resources address and access time and, the SOT state of termination include network environment information and Charged state；

And web-based history access record and the SOT state of termination are pre-processed, obtain state set and observation set, packet It includes：

It is corresponded to according to the program number of the preset rules matching application program for accessing network and the network resources address Network numbering；

Working day and day off are distinguished according to the access time, respectively working day and day off assigns different date codes；

Period according to belonging to the access time determines time domain, wherein, preset time section in consecutive days is carried out Respectively obtain the period, the period and time domain associated storage；

Judge whether terminal accesses wireless network according to the network environment information, network-like state value is determined according to judging result；

It is numbered according to described program, the conjunction of date codes, time domain, network-like state value and state-of-charge value generating states set, and Observation set is formed according to the network numbering.

6. a kind of Internet resources pre-add support method, which is characterized in that including：

Current network access record is obtained according to the setting period and records the associated SOT state of termination with current network access；

By the access prediction model that the current network accesses record and SOT state of termination input is built in advance, wherein, the access Prediction model is the model that corresponding SOT state of termination training is recorded according to web-based history access record and web-based history access；

The network resource identifier to be visited of the access prediction model output is obtained, it is corresponding to download the network resource identifier Network data, and it is stored in local cache.

7. according to the method described in claim 6, it is characterized in that, the current network is accessed into record and SOT state of termination input The access prediction model built in advance, including：

The access prediction model that current state input is built in advance, with lower for the moment by the access prediction model prediction Corresponding NextState is carved, and determines the observation probability of the corresponding network resources address of the NextState respectively.

8. according to the method described in claim 6, it is characterized in that, downloading the corresponding network data of the network resource identifier Before, it further includes：

Judge whether terminal accesses wireless network；

When the status information meets preset condition, the behaviour for downloading the corresponding network data of the network resource identifier is performed Make；

And when the status information is unsatisfactory for preset condition, the corresponding network data of the network resource identifier is downloaded, is wrapped It includes：

Target network data are determined according to the data type of the corresponding network data of the network resource identifier and/or data volume, The target network data are loaded.

9. according to the method described in claim 6, it is characterized in that, downloading the corresponding network data of the network resource identifier Later, it further includes：

The second network that the first network resource address that the network resource identifier includes is linked to by web crawlers acquisition provides Source address loads the corresponding network data of second network resources address, and is stored in local cache.

10. it according to the method described in claim 6, it is characterized in that, further includes：

Obtain the memory space that network data occupies in the local cache；

When the memory space is more than given threshold, determined according to the cache-time of the network data and/or frequency of use Data to be cleaned, and remove the data to be cleaned.

11. according to the method described in any one of claim 6-10, which is characterized in that further include：

Obtain the network access request of application program；

The local cache is inquired according to the network access request, determines the corresponding network data of the network access request Preload condition；

If the corresponding network data of the network access request has preloaded, the network data is read, and is back to described Application program；

If the corresponding network data of the network access request does not preload, the network data is downloaded by outer net.

12. a kind of model construction device, which is characterized in that including：

Historical record acquisition module, for obtaining web-based history access record and recording associated end with web-based history access End state；

Matrix deciding module, for according to the web-based history access record and the SOT state of termination determine state transition probability matrix and Observation probability matrix；

Model construction module accesses prediction model for being based on the state transition probability matrix and observation probability matrix structure.

13. a kind of Internet resources pre-load means, which is characterized in that including：

Current record acquisition module, for obtaining current network access record according to the setting period and being accessed with the current network Record the associated SOT state of termination；

Mould is predicted in mode input module, the access for current network access record and SOT state of termination input to be built in advance Type, wherein, the access prediction model is to record corresponding end according to web-based history access record and web-based history access The model of end state training；

Preloaded components for obtaining the network resource identifier to be visited of the access prediction model output, download the net The corresponding network data of network resource identification, and it is stored in local cache.

14. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is by processor The model building method as described in any in claim 1 to 5 is realized during execution；Or the program is realized when being executed by processor Internet resources pre-add support method as described in any in claim 6 to 11.

15. a kind of terminal including memory, processor and stores the computer journey that can be run on a memory and on a processor Sequence, which is characterized in that the processor realizes the mould as described in any in claim 1 to 5 when performing the computer program Type construction method；Alternatively, the processor is realized when performing the computer program as described in any in claim 6 to 11 Internet resources pre-add support method.