CN106021391A - Product comment information real-time collection method based on Storm - Google Patents
Product comment information real-time collection method based on Storm Download PDFInfo
- Publication number
- CN106021391A CN106021391A CN201610313091.3A CN201610313091A CN106021391A CN 106021391 A CN106021391 A CN 106021391A CN 201610313091 A CN201610313091 A CN 201610313091A CN 106021391 A CN106021391 A CN 106021391A
- Authority
- CN
- China
- Prior art keywords
- product
- module
- time
- review information
- queue
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a product comment information real-time collection method based on Storm. The method is carried out in a collection system based on a Storm platform. The method comprises the steps that a product capturing module captures data from a network periodically; a preprocessing module initializes attribute parameters of products according to product Ids; a scheduling module packages the data into Tuple and transmits the Tuple to a Storm cluster; an update detection module detects whether new comments are generated or not and sends the new comments to a comment information capturing module; the comment information capturing module detects the products with updated comment information and carries out distributed capturing; an access interval adjusting module dynamically adjusts the next collection time of the products; and so on. According to the method, a traditional web crawler and an open source distributed stream processing frame Storm are combined; therefore, the traditional web crawler can be operated in the stream processing platform; the real-time information collection performance is greatly improved; and the high practical value and practical significance are realized.
Description
Technical field
The invention belongs to network information processing technical field, be specifically related to product review information based on Storm and adopt in real time
Diversity method.
Background technology
Along with increasing rapidly of Internet user's quantity, network increasingly become people life in the most retrievable one
Point, the particularly development of ecommerce, change people's consumption pattern so that the shopping of people is very convenient, quick.Consumer
Will not be dejected for alone people goes window-shopping again, no longer do shopping and worried, less with in order to choose one for not free going window-shopping
Commodity and east run west run.Under the consumption mode of ecommerce, it is only necessary to a bank card, a computer, consumer can buy
To any thing, it is not necessary to a lot of time, it is not necessary to a lot of labours.In order to provide, to consumer, body of preferably doing shopping
Testing, most of e-commerce websites are that consumer provides the platform of mutually exchange to deliver its commenting for certain part product or service
Opinion, as a kind of novel oral marketing mode, the review information of product has become as user and understands product quality and service
Important information source, it is clear that the decision-making in purchasing of consumer can be produced a very large impact by the quality of review information.Additionally, for businessman
Process the negative reviews information of commodity timely and solve the problem of consumer feedback, to the favorable image and the commodity that keep businessman
Sales volume be particularly important.Therefore, the review information of product in acquisition electricity business website the most as early as possible, and then real-time assurance
Commodity public sentiment, all seems particularly significant to businessman or consumer.
But, in the last few years, the pertinent literature of real-time information collection is also the most many, but is requiring that premise is gone down pre-accurately
Surveying product and commenting on time of generation next time is a highly difficult task, and the most good forecast model and algorithm do not go out
Existing.But have a lot to the research field capturing webpage relevant in real time, and such as predict the renewal frequency of webpage, method is roughly divided into
The most several classes:
(1) regeneration behavior of webpage is simulated based on mathematical modeies such as Poisson distributions, proposing to grinding of experiencing of this thought
Study carefully the time the most long, but this is difficult to have substantial progress and breakthrough, secondly, the only time prediction to coarseness of this model
Effect is preferable, and poor for fine-grained time prediction effectiveness comparison.
(2) behavior of homepage modification is simulated with a mathematics probabilistic model, with this mathematical model Approximate prediction webpage
Amendment frequency, but it not the most the highest with the accuracy rate of webpage behavior matching and stability in this way.
(3) A.K.Sharma et al. proposes the thought that comparison is new, it is not necessary to the historical record that webpage is complete, accesses every time
The next update time of corresponding web page all can be adjusted after webpage, and according to constantly capturing the web-page histories amendment frequency accumulated
The data of rate update the time next time and are predicted webpage, and weak point is excessively to pay close attention to the renewal frequency of webpage, and
Capture the inconsiderate of the new webpage aspect details revised consideration in time.
Summary of the invention
It is an object of the invention to provide a kind of product review information real-time collecting method based on Storm, real to improve
Time gather accuracy rate and stability.
For achieving the above object, the present invention is by the following technical solutions:
The invention discloses product review information real-time collecting method based on Storm, the method is based on Storm platform
Acquisition system carry out, described acquisition system include product handling module, pretreatment module, scheduler module, renewal detection mould
Block, review information handling module, HBase data memory module, access interval adjusting module；The method includes:
A. product handling module is activated by Time Triggered server, periodically captures data from network, obtains product
Property parameters and be deposited in HBase data memory module, check that the product attribute Id newly grabbed has deposited simultaneously
In queue to be climbed, if existed, ignoring, being otherwise sent to product pretreatment module and carrying out pretreatment.
B. pretreatment module initializes the property parameters of product according to product I d, and product I d after initializing is through url
Link optimized puts into queue fetcher_queue to be climbed after processing.
C. scheduler module constantly reads data from fetcher_queue queue, and encapsulates data into Tuple transmitting
In Storm cluster.
D. update the new comment of detection module check whether there is to produce, if this product has newly comments on generation, just sent
To review information handling module.
E. to updating detection module, review information handling module detects that the product that review information updates carries out distributed
Capture.
F. access interval adjusting module and dynamically adjust the time that product gathers next time, be then reentered into team to be climbed
In row fetcher_queue.
Further, the described interval adjusting module that accesses adjusts the time that product gathers next time, comprises the following steps:
S1. product history minimum renewal frequency f is obtainedl, maximum renewal frequency fuAnd acquisition interval interval.
S2. the times of collection n that product is current is obtainedt, review information generation update times nc, calculate product current more
New frequency fc=nc/nt。
S3. acquisition interval interval of product is adjusted according to following equation,
Next_interval=interval+ Δ t (formula 1)
Wherein μ (x): be unit-step function,
S4. current comment amount n of product, time interval t of the last review information generation are passed through1And it is the last
The t maximum lag time of the product review information gathered2As its priority quantizing factor, respectively to n, t1、t2Do 0～1 interval
Normalization operation, be calculated as follows this product priority p now, be reentered into queue fetcher_queue to be climbed；
Preferably, the step that pretreatment module carries out pretreatment to product is as follows:
S1. obtain the current comment sum of product and the last comment time, comment according to comment sum and the last time
Opinion initializes its acquisition interval interval.
S2. maximum renewal frequency f of product is initializeduWith minimum renewal frequency fl, by product I d through url link optimized
Module processes, and then the acquisition interval of this product and the url link optimized is put into frequency initialization queue
frequency_init_queue；As long as frequency_init_queue is not empty, frequency initializes reptile freCrawer i.e.
With the acquisition interval of this product, this product is carried out the collection of 500 times, and records the frequency n occurring to update altogetherc, collection terminates
After, calculate its renewal frequency fc=nc/ 500, and use fcInitialize maximum renewal frequency f of productuWith minimum renewal frequency fl。
S3. product I d after pretreated is put into queue fetcher_ to be climbed after url link optimized resume module
queue。
Preferably, update detection module and obtain the up-to-date comment sum current_total of product, then with the going through of product
The total old_total contrast of commentary on historical events or historical records opinion, if greater than historical review sum, then the historical review of upgrading products is total, it is secondary to gather
Number, review information generation update times and the last acquisition time, and this product is sent to review information handling module
Carry out review information collection, the otherwise times of collection of a upgrading products and the last acquisition time.
Preferably, the step of the described distributed crawl of review information handling module is as follows:
S1. the last comment time recent_comments_time of this product is obtained；
S2. according to the result of the detection updating detection module, the review information of corresponding number of pages is downloaded；
S3. the review information downloaded further is cleaned, found out the last comment time that product is made, and
Update the comment time this product the last time；
S4. the new product review information captured is sent to HBase data memory module store.
Use after technique scheme, the beneficial effects of the present invention is: the present invention by traditional web crawlers with increase income
Distributed stream processes framework-Storm and combines, and makes full use of that Storm is distributed, Error Tolerance, no data are lost, can be expanded
Malleability is strong and the advantage such as real-time process, and by dynamically adjusting acquisition time interval arithmetic and collection scheduling strategy respectively to adopting
Collection time point and collection scheduling order are further optimized, and have the most both decreased the network request number of times of reptile, but also have made
Ask more rationally, effectively, finally by the handling capacity of test analysis Storm cluster with process the performances such as delay, result every time
Demonstrate Storm can be competent at the network data acquisition task under requirement of real-time high field scape.
Accompanying drawing explanation
Fig. 1 is process structure figure of the present invention.
Fig. 2 is review information of the present invention distributed crawl Organization Chart.
Specific embodiments
In order to make the purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, right
The present invention is further elaborated.
As it is shown in figure 1, the invention discloses product review information real-time collecting method based on Storm.The inventive method
Carrying out in acquisition system based on Storm platform, acquisition system includes product handling module, pretreatment module, scheduler module, more
New detection module, review information handling module, HBase data memory module, access interval adjusting module, enter this system below
Row review information real-time collecting method is described in detail:
A. product handling module
Activated by Time Triggered server, from network, periodically capture data, obtain the property parameters of product and deposit
Enter in HBase data memory module, check that the product attribute Id newly grabbed has existed queue to be climbed simultaneously, if
Exist, ignored, be otherwise sent to product pretreatment module and carry out pretreatment.
B. pretreatment module
Initialize the property parameters of product according to product I d, product I d after initializing processes through url link optimized
After to put into queue fetcher_queue to be climbed, queue fetcher_queue to be climbed be one in Beanstalkd queue
Pipeline.Specifically comprise the following steps that
S1. obtain the current comment sum of product and the last comment time, comment according to comment sum and the last time
Opinion initializes its acquisition interval interval.
S2. maximum renewal frequency f of product is initializeduWith minimum renewal frequency fl, by product I d through url link optimized
Module processes, and then the acquisition interval of this product and the url link optimized is put into frequency initialization queue
frequency_init_queue；As long as frequency_init_queue is not empty, frequency initializes reptile freCrawer i.e.
With the acquisition interval of this product, this product is carried out the collection of 500 times, and records the frequency n occurring to update altogetherc, collection terminates
After, calculate its renewal frequency fc=nc/ 500, and use fcInitialize maximum renewal frequency f of productuWith minimum renewal frequency fl。
S3. product I d after pretreated is put into queue fetcher_ to be climbed after url link optimized resume module
queue。
C. scheduler module
This module is mainly responsible for the acquisition order of product in acquisition system.Owing to this acquisition system is based on Storm platform
, and Storm is as real-time stream processing platform, provides data for it the most endlessly, it could normally work,
Spout is as the core component of Storm, and it is responsible in Storm cluster launching data stream, and therefore, this module can continuous quilt
Spout calls, thus constantly reads data from fetcher_queue queue, and encapsulates data into Tuple and be transmitted into
In Storm cluster.
D. detection module is updated
This module is primarily used to detect whether product has new comment to produce.It comes according to some acquisition parameters of product
Judge whether product has new comment to produce, if this product has newly comments on generation, be just sent to review information handling module:
Update detection module and obtain the up-to-date comment sum current_total of product, then with the historical review sum old_ of product
Total contrasts, and if greater than historical review sum, then the historical review sum of upgrading products, times of collection, review information occur
Update times and the last acquisition time, and this product is sent to review information handling module carries out review information and adopt
Collection, the otherwise times of collection of a upgrading products and the last acquisition time.
E. review information handling module
This module is the core of whole acquisition system.This module detects what review information updated to updating detection module
Product carries out distributed crawl.Owing to it operates in Storm cluster, it is thereby achieved that the distribution to product review information
Formula captures, distributed reptile Organization Chart as in figure 2 it is shown, in Organization Chart review information handling module comprise following assembly:
(1) the control node in Nimbus:Storm cluster, is responsible for resource distribution and task scheduling in Storm cluster.
(2) working node in Supervisor:Storm cluster, is responsible for accepting the task of Nimbus distribution, starts and stop
The worker process only to one's name managed.
(3) Worker:Storm cluster runs the concrete process processing logic.
(4) Executor: the task of same spout/bolt may share a physical thread, and this thread is referred to as
Executor。
(5) in Task:worker, the thread of each spout/bolt is referred to as a task.
(6) Zookeeper: be responsible between nimbus and supervisor the service coordinated.
(7) HBase: be responsible for the storage of product review information.
The step of the distributed crawl of review information handling module is as follows:
S1. the last comment time recent_comments_time of this product is obtained.
S2. according to the result of the detection updating detection module, the review information of corresponding number of pages is downloaded.
S3. the review information downloaded further is cleaned, found out the last comment time that product is made, and
Update the comment time this product the last time.
S4. the new product review information captured is sent to HBase data memory module store.
F. interval adjusting module is accessed
This module dynamically adjusts the time that product gathers next time, is then reentered into queue fetcher_ to be climbed
In queue.Comprise the following steps:
S1. product history minimum renewal frequency f is obtainedl, maximum renewal frequency fuAnd acquisition interval interval.
S2. the times of collection n that product is current is obtainedt, review information generation update times nc, calculate product current more
New frequency fc=nc/nt。
S3. acquisition interval interval of product is adjusted according to formula 1, formula 2.
Next_interval=interval+ Δ t (formula 1)
Wherein μ (x): be unit-step function,
Analyzed by formula 1 above, formula 2 it appeared that the adjustment to product acquisition interval specifically can be divided into following three
The situation of kind:
(1) if. the current review information calculated produces frequency fcBetween lower bound flWith upper bound fuBetween, it is not necessary to adjust, this
Time Δ t=0.
(2) if. the current review information calculated produces frequency fcLess than lower bound fl, illustrate that product produces review information
Frequency downshift, needs to increase compared with history acquisition interval, now Δ t > 0, and resets lower bound flCollection with product
Interval.
(3) if. the current review information calculated produces frequency fcMore than upper bound fu, illustrate that product produces review information
Frequency accelerates, and needs to reduce, now Δ t < 0, and reset upper bound f compared with history acquisition intervaluCollection with product
Interval.
S4. current comment amount n of product, time interval t of the last review information generation are passed through1And it is the last
The t maximum lag time of the product review information gathered2As its priority quantizing factor, respectively to n, t1、t2Do 0～1 interval
Normalization operation, calculate this product priority p now by formula 3, be reentered into queue fetcher_queue to be climbed.
E.HBase data memory module
This module mainly coordinates in the operation of each module and stores the data being resolved to.The relationship type number that data are traditional
Having shown mass data storage weak according to storehouse, therefore, the present invention uses novel non-relational database HBase to store
The base attribute of product and the review information of product.
To sum up, present invention product review information based on Storm real-time collecting method by traditional web crawlers with increase income
Distributed stream processes framework-Storm and combines, and dynamically adjusts acquisition time interval and collection scheduling strategy, improves and adopt in real time
The accuracy rate of collection and stability.
The above, the only present invention preferably detailed description of the invention, but protection scope of the present invention is not limited thereto,
Any those familiar with the art in the technical scope that the invention discloses, the change that can readily occur in or replacement,
All should contain within protection scope of the present invention.
Claims (5)
1. product review information real-time collecting method based on Storm, it is characterised in that: the method is based on Storm platform
Acquisition system is carried out, described acquisition system include product handling module, pretreatment module, scheduler module, renewal detection module,
Review information handling module, HBase data memory module, access interval adjusting module；The method includes:
A. product handling module is activated by Time Triggered server, periodically captures data from network, obtains the genus of product
Property parameter is also deposited in HBase data memory module, checks that the product attribute Id newly grabbed has existed simultaneously and treats
Climbing queue, if existed, ignoring, be otherwise sent to product pretreatment module and carry out pretreatment；
B. pretreatment module initializes the property parameters of product according to product I d, and product I d after initializing links through url
Optimization puts into queue fetcher_queue to be climbed after processing；
C. scheduler module constantly reads data from fetcher_queue queue, and encapsulates data into Tuple and be transmitted into
In Storm cluster；
D. update the new comment of detection module check whether there is to produce, if this product has newly comments on generation, be just sent to comment
Opinion information scratching module；
E. to updating detection module, review information handling module detects that the product that review information updates carries out distributed crawl；
F. access interval adjusting module and dynamically adjust the time that product gathers next time, be then reentered into queue to be climbed
In fetcher_queue.
2. product review information real-time collecting method based on Storm as claimed in claim 1, it is characterised in that: described
Access interval adjusting module and adjust the time that product gathers next time, comprise the following steps:
S1. product history minimum renewal frequency f is obtainedl, maximum renewal frequency fuAnd acquisition interval interval；
S2. the times of collection n that product is current is obtainedt, review information generation update times nc, calculate the renewal frequency that product is current
Rate fc=nc/nt；
S3. acquisition interval interval of product is adjusted according to following equation,
Next_interval=interval+ Δ t (formula 1)
Wherein μ (x): be unit-step function,
S4. current comment amount n of product, time interval t of the last review information generation are passed through1And the last time gathers
The t maximum lag time of product review information2As its priority quantizing factor, respectively to n, t1、t2Do returning of 0～1 interval
One changes operation, is calculated as follows this product priority p now, is reentered into queue fetcher_queue to be climbed；
3. product review information real-time collecting method based on Storm as claimed in claim 2, it is characterised in that pretreatment
The step that module carries out pretreatment to product is as follows:
S1. the current comment sum of product and the last comment time are obtained, at the beginning of comment sum and the last comment
Its acquisition interval interval of beginningization；
S2. maximum renewal frequency f of product is initializeduWith minimum renewal frequency fl, by product I d through url link optimized module
Process, then the acquisition interval of this product and the url link optimized are put into frequency initialization queue frequency_
init_queue；As long as frequency_init_queue is not empty, frequency initializes reptile freCrawer i.e. with this product
Acquisition interval carries out the collection of 500 times to this product, and records the frequency n occurring to update altogetherc, gather after terminating, calculate it more
New frequency fc=nc/ 500, and use fcInitialize maximum renewal frequency f of productuWith minimum renewal frequency fl；
S3. product I d after pretreated is put into queue fetcher_ to be climbed after url link optimized resume module
queue。
4. product review information real-time collecting method based on Storm as claimed in claim 3, it is characterised in that update inspection
Survey module and obtain the up-to-date comment sum current_total of product, then with historical review sum old_total pair of product
Ratio, if greater than historical review sum, then there is renewal time in the historical review sum of upgrading products, times of collection, review information
Several and the last acquisition time, and this product is sent to review information handling module carries out review information collection, otherwise
The times of collection of upgrading products and the last acquisition time.
5. product review information real-time collecting method based on Storm as claimed in claim 4, it is characterised in that described
The step of the distributed crawl of review information handling module is as follows:
S1. the last comment time recent_comments_time of this product is obtained；
S2. according to the result of the detection updating detection module, the review information of corresponding number of pages is downloaded；
S3. the review information downloaded further is cleaned, find out the last comment time that product is made, and update
The comment time this product the last time；
S4. the new product review information captured is sent to HBase data memory module store.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610313091.3A CN106021391B (en) | 2016-05-11 | 2016-05-11 | Product review information real-time collecting method based on Storm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610313091.3A CN106021391B (en) | 2016-05-11 | 2016-05-11 | Product review information real-time collecting method based on Storm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106021391A true CN106021391A (en) | 2016-10-12 |
| CN106021391B CN106021391B (en) | 2019-06-21 |
Family
ID=57100199
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610313091.3A Active CN106021391B (en) | 2016-05-11 | 2016-05-11 | Product review information real-time collecting method based on Storm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106021391B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107169024A (en) * | 2017-04-11 | 2017-09-15 | 微梦创科网络科技（中国）有限公司 | The operation system and service implementation method of a kind of compatible type |
| CN108614841A (en) * | 2016-12-13 | 2018-10-02 | 北京国双科技有限公司 | The method of adjustment and device of time interval |
| CN109388748A (en) * | 2018-09-26 | 2019-02-26 | 深圳壹账通智能科技有限公司 | A kind of answering method of comment information, storage medium and server |
| CN109857795A (en) * | 2019-01-02 | 2019-06-07 | 拉卡拉支付股份有限公司 | A kind of tables of data cut-in method and system based on prediction model |
| CN113312526A (en) * | 2021-06-29 | 2021-08-27 | 平安资产管理有限责任公司 | Network information dynamic acquisition method and device, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207855A (en) * | 2013-04-12 | 2013-07-17 | 广东工业大学 | Fine-grained sentiment analysis system and method specific to product comment information |
| CN103678564A (en) * | 2013-12-09 | 2014-03-26 | 国家计算机网络与信息安全管理中心 | Internet product research system based on data mining |
| US20160042035A1 (en) * | 2014-08-08 | 2016-02-11 | International Business Machines Corporation | Enhancing textual searches with executables |
-
2016
- 2016-05-11 CN CN201610313091.3A patent/CN106021391B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207855A (en) * | 2013-04-12 | 2013-07-17 | 广东工业大学 | Fine-grained sentiment analysis system and method specific to product comment information |
| CN103678564A (en) * | 2013-12-09 | 2014-03-26 | 国家计算机网络与信息安全管理中心 | Internet product research system based on data mining |
| US20160042035A1 (en) * | 2014-08-08 | 2016-02-11 | International Business Machines Corporation | Enhancing textual searches with executables |
Non-Patent Citations (2)
| Title |
|---|
| NIRAJ SINGHAL,ASHUTOSH DIXITDR.A.K.SHARMA: "Design of a priority based frequency regulated incremental crawler", 《2010 INTERNATIONAL JOURNAL OF COMPUTER APPLICATIONS》 * |
| 付志鸿: "基于Storm云平台的分布式网络爬虫技术研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108614841A (en) * | 2016-12-13 | 2018-10-02 | 北京国双科技有限公司 | The method of adjustment and device of time interval |
| CN108614841B (en) * | 2016-12-13 | 2021-09-07 | 北京国双科技有限公司 | Time interval adjusting method and device |
| CN107169024A (en) * | 2017-04-11 | 2017-09-15 | 微梦创科网络科技（中国）有限公司 | The operation system and service implementation method of a kind of compatible type |
| CN109388748A (en) * | 2018-09-26 | 2019-02-26 | 深圳壹账通智能科技有限公司 | A kind of answering method of comment information, storage medium and server |
| WO2020062672A1 (en) * | 2018-09-26 | 2020-04-02 | 深圳壹账通智能科技有限公司 | Method for replying to comment information, storage medium, server and device |
| CN109857795A (en) * | 2019-01-02 | 2019-06-07 | 拉卡拉支付股份有限公司 | A kind of tables of data cut-in method and system based on prediction model |
| CN113312526A (en) * | 2021-06-29 | 2021-08-27 | 平安资产管理有限责任公司 | Network information dynamic acquisition method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106021391B (en) | 2019-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106021391A (en) | Product comment information real-time collection method based on Storm | |
| CN103761309B (en) | Operation data processing method and system | |
| CN106502772A (en) | Electric quantity data batch high speed processing method and system based on distributed off-line technology | |
| CN106296305A (en) | Electric business website real-time recommendation System and method under big data environment | |
| CN104090894B (en) | The method of online parallel computation recommendation information, device and server | |
| Tsuchiya et al. | Big data processing in cloud environments | |
| CN109961204A (en) | Quality of service analysis method and system under a kind of micro services framework | |
| CN108197737A (en) | A kind of method and system for establishing medical insurance hospitalization cost prediction model | |
| Kuang et al. | Personalized services recommendation based on context-aware QoS prediction | |
| Zhang et al. | Urban traffic flow forecast based on FastGCRNN | |
| Flouris et al. | Ferari: A prototype for complex event processing over streaming multi-cloud platforms | |
| CN109087030A (en) | Realize method, General Mobile crowdsourcing server and the system of the crowdsourcing of C2C General Mobile | |
| CN115860529A (en) | Supply chain carbon accounting system based on industrial internet | |
| CN108833294B (en) | Low-bandwidth-overhead flow scheduling method for data center wide area network | |
| Wang et al. | Big data in telecommunication operators: data, platform and practices | |
| CN103412903A (en) | Method and system for interested object prediction based real-time search of Internet of Things | |
| CN106202383A (en) | A kind of network bandwidth accounting dynamic prediction method being applied to web crawlers and system | |
| Paul et al. | A comprehensive review of green computing: Past, present, and future research | |
| Angerd et al. | Distributed training of graph convolutional networks using subgraph approximation | |
| CN110348928A (en) | Information-pushing method, device and computer readable storage medium | |
| CN111049898A (en) | Method and system for realizing cross-domain architecture of computing cluster resources | |
| Li et al. | Prediction of china’s housing price based on a novel grey seasonal model | |
| Siregar et al. | Big Data Analytics Based Model for Red Chili Agriculture in Indonesia | |
| Alifah et al. | Smart monitoring of rice logistic employing internet of things network | |
| CN101741624A (en) | Internet composite service performance fault-tolerant system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |