CN106991182A - A kind of friends recognition methods based on high frequency brushing card data - Google Patents
A kind of friends recognition methods based on high frequency brushing card data Download PDFInfo
- Publication number
- CN106991182A CN106991182A CN201710226852.6A CN201710226852A CN106991182A CN 106991182 A CN106991182 A CN 106991182A CN 201710226852 A CN201710226852 A CN 201710226852A CN 106991182 A CN106991182 A CN 106991182A
- Authority
- CN
- China
- Prior art keywords
- data
- similar
- similar behavior
- card
- friends
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/288—Entity relationship models
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)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本发明是一种基于高频刷卡数据的朋友关系识别方法,主要包括数据预处理、相似行为矩阵的生成、相似行为阈值的确定和朋友关系的识别;1)数据预处理;获取当前所有用户的高频刷卡数据,提取数据中有效信息,存入统一的数据仓库中;2)相似行为矩阵的生成;3)相似行为阈值的确定;提取相似行为矩阵对角线上方的所有元素,生成帕累托图,取累计百分比首次超过α,0<α<100%所对应的相似行为的次数作为相似行为阈值,α通常取90%;4)朋友关系的识别;找出所有相似行为超过相似行为阈值的用户对,将他们确认为朋友关系,同时输出所有的朋友对。本发明能够利用用户产生的高频刷卡数据有效识别用户在现实生活中的朋友关系。The present invention is a friend relationship identification method based on high-frequency card swiping data, which mainly includes data preprocessing, generation of similar behavior matrix, determination of similar behavior threshold and identification of friend relationship; 1) data preprocessing; High-frequency card swiping data, extract valid information in the data, and store in a unified data warehouse; 2) Generation of similar behavior matrix; 3) Determination of similar behavior threshold; extract all elements above the diagonal of the similar behavior matrix, and generate Pare Totu, take the cumulative percentage exceeding α for the first time, and the number of similar behaviors corresponding to 0<α<100% is used as the similar behavior threshold, and α is usually taken as 90%; 4) Recognition of friendship; find out all similar behaviors exceeding the similar behavior threshold User pairs, confirm them as friends, and output all friend pairs at the same time. The invention can utilize the high-frequency swiping card data generated by the user to effectively identify the friendship relationship of the user in real life.
Description
技术领域technical field
本发明属于信息处理和特征识别领域,涉及到一种基于高频刷卡数据的朋友关系识别方法。The invention belongs to the fields of information processing and feature recognition, and relates to a friend relationship recognition method based on high-frequency card swiping data.
背景技术Background technique
现在如果我们假设你从一个集体到了另一个新的集体中,那么你会有多大的可能性和你周围的某个人成为朋友呢?又从社会学的方面而言,在一个群体中,两两之间的朋友关系又能占到多大比例呢?要回答这些问题,首先就要建立一种稳定、高效同时具有大规模性质的朋友识别方法。另一方面,从商业角度来说,个人社会朋友关系的识别对一些企业或公司拓展新用户,以及对老用户开展个性化服务都具有十分重要的意义。Now if we assume that you move from one group to another new group, how likely are you to become friends with someone around you? From a sociological point of view, in a group, what proportion can the friendship between two pairs account for? To answer these questions, we must first establish a stable, efficient and large-scale friend identification method. On the other hand, from a business point of view, the identification of personal social friend relationships is of great significance for some enterprises or companies to expand new users and provide personalized services to old users.
近年来,随着互联网的普及和发展,越来越多的专家学者乐于用一些社交软件,网络平台,虚拟游戏设备来识别朋友关系。虽然这些方法对社会研究,企业发展能产生一定的指导性意义,但是基于这些方法,认真分析可发现三点明显的不足:一是用户在填写注册信息时往往会存在虚假成分,不能真实反映自身的信息;二是用户的网上行为并不能真实的反映人们现实世界的朋友关系,比如银行、超市、健身房等服务性场所,却只有现实世界的朋友关系才能帮助它们有效拓展客户源;三是在网络虚拟平台上得到的某些朋友关系及其薄弱,或者地域上相差及远,很明显这些所谓的朋友关系或者不存在,或者不处在一个特定的群体里,对我们的研究而言没有太大的实际意义。In recent years, with the popularization and development of the Internet, more and more experts and scholars are willing to use some social software, network platforms, and virtual game devices to identify friend relationships. Although these methods can have certain guiding significance for social research and enterprise development, based on these methods, careful analysis can reveal three obvious deficiencies: First, users often have false elements when filling in registration information, which cannot truly reflect themselves Second, users’ online behavior cannot truly reflect people’s real-world friendships, such as banks, supermarkets, gyms and other service places, but only real-world friendships can help them effectively expand their customer base; Some friendships obtained on the virtual network platform are extremely weak, or the regions are far apart. It is obvious that these so-called friendships either do not exist, or are not in a specific group, which is not too important for our research. great practical significance.
当前的朋友识别系统通常是基于用户的注册信息,以及用户的上网行为来挖掘用户之间的朋友关系,对互联网技术有了应用却又恰恰忽略了现实社会,人们真实生活记录的重要性,反倒显得得不偿失。经过认真分析研究发现随着现代信息技术的快速发展和网络技术的不断成熟,以及近些年来大数据的分析应用技术已经进入人们生活的方方面面,记录人们日常真实行为一卡通作为信息技术的产物,对于改进人们生活的便捷程度和推动数字化管理起到了巨大的作用。由一卡通产生的反映人们行为的高频数据能够有效识别人与人之间关系,那些在大量的数据中拥有着相似刷卡行为的个体亲密度更高,更有可能是朋友关系。由此,产生了一种基于高频刷卡数据的朋友关系识别方法。The current friend recognition system is usually based on the user's registration information and the user's online behavior to mine the friend relationship between users. With the application of Internet technology, it just ignores the importance of people's real life records in the real society. It seems that the loss outweighs the gain. After careful analysis and research, it is found that with the rapid development of modern information technology and the continuous maturity of network technology, as well as the analysis and application technology of big data in recent years, it has entered all aspects of people's lives. As a product of information technology, the one-card that records people's daily real behavior is of great importance. Improving the convenience of people's lives and promoting digital management have played a huge role. The high-frequency data reflecting people's behavior generated by the One Card can effectively identify the relationship between people. Those individuals who have similar card swiping behaviors in a large amount of data have higher intimacy and are more likely to be friends. Thus, a friend relationship recognition method based on high-frequency card swiping data is produced.
日常生活中,高频刷卡数据的来源极其广泛。例如,高校学生一日三餐的刷卡数据,以及进出校门,宿舍,图书馆借书、自习,超市购物等数据都可以作为很好的研究样本。并且这种大数据的特点从一定程度上保证了研究结果的稳定性和可靠性。另一方面,这些大数据可以利用先进的计算机技术进行分析处理,很好的保证了研究的准确性和高效性。In daily life, the sources of high-frequency card swiping data are extremely wide. For example, the data of swiping credit cards of college students for three meals a day, as well as the data of entering and leaving the school gate, dormitory, library borrowing, self-study, supermarket shopping, etc. can be used as good research samples. And the characteristics of this big data guarantee the stability and reliability of the research results to a certain extent. On the other hand, these big data can be analyzed and processed by advanced computer technology, which ensures the accuracy and efficiency of the research.
运用这些高频刷卡数据和先进的应用软件以及计算分析技术真正实现了从真实的日常生活中来识别朋友关系。从研究者的角度来说,朋友是从各种各样的生活数据中得来,它的真实性和可靠性无须赘述,对进一步的社会关系的发展和演变的研究,以及其规律的探寻有着不可代替的重要作用。Using these high-frequency card swiping data and advanced application software and computing analysis technology has truly realized the identification of friendship from real daily life. From a researcher's point of view, friends are obtained from various life data, and its authenticity and reliability need not be repeated. It is of great significance to further research on the development and evolution of social relations and the exploration of its laws. irreplaceable important role.
发明内容Contents of the invention
针对当前朋友关系识别方法中存在的仅仅依靠用户注册信息和用户上网行为进行朋友关系识别的不足,本发明提出了一种基本高频刷卡数据的朋友关系识别方法。以满足商家有效挖掘新用户和对老用户进行个性化服务的需求。Aiming at the shortcomings in current friend relationship identification methods that only rely on user registration information and user online behaviors for friend relationship identification, the present invention proposes a friend relationship identification method based on high-frequency card swiping data. To meet the needs of merchants to effectively tap new users and provide personalized services to old users.
本发明采用的技术方案是:The technical scheme adopted in the present invention is:
一种基于高频刷卡数据的朋友关系识别方法,主要包括数据预处理、相似行为矩阵的生成、相似行为阈值的确定和朋友关系的识别;A friend relationship identification method based on high-frequency card swiping data, mainly including data preprocessing, generation of similar behavior matrix, determination of similar behavior threshold and identification of friend relationship;
1)数据预处理;获取当前所有用户的高频刷卡数据,提取数据中有效信息,包括用户的ID,刷卡的地点,刷卡的时间,刷卡操作的类型,删除数据中所有的标点,仅保留数字信息,并对数据进行统一的命名,存入统一的数据仓库中;1) Data preprocessing; obtain high-frequency card swiping data of all current users, extract valid information in the data, including user ID, location of swiping card, time of swiping card, type of card swiping operation, delete all punctuation in the data, and only keep numbers information, and uniformly name the data, and store them in a unified data warehouse;
2)相似行为矩阵的生成;2) Generation of similar behavior matrix;
(a)对于数据仓库中的数据而言,当任意两个用户的刷卡记录的地点一致,同时刷卡时间间隔不超过t秒时,则称这两个用户之间存在一次相似行为;(a) For the data in the data warehouse, when any two users have the same card swiping records at the same location, and the time interval between swiping cards at the same time does not exceed t seconds, it is said that there is a similar behavior between the two users;
(b)设定时间区间T,提取数据仓库中在时间区间T范围内的数据;在时间区间T范围内,用户m和用户n的相似行为次数就构成了相似行为矩阵的一个元素am,n;相似行为矩阵为主对角元素均为0的对称矩阵;(b) Set the time interval T, and extract the data within the time interval T in the data warehouse; within the time interval T, the number of similar behaviors of user m and user n constitutes an element a m of the similar behavior matrix, n ; The similar behavior matrix is a symmetric matrix whose main and diagonal elements are all 0;
3)相似行为阈值的确定;提取相似行为矩阵对角线上方的所有元素,生成帕累托图,取累计百分比首次超过α,0<α<100%所对应的相似行为的次数作为相似行为阈值,α通常取90%;3) Determination of the similar behavior threshold; extract all elements above the diagonal of the similar behavior matrix, generate a Pareto diagram, and take the number of similar behaviors corresponding to the cumulative percentage exceeding α for the first time, 0<α<100% as the similar behavior threshold , α usually takes 90%;
4)朋友关系的识别;找出所有相似行为超过相似行为阈值的用户对,将他们确认为朋友关系,同时输出所有的朋友对。4) Recognition of friend relationship; find out all user pairs whose similar behavior exceeds the similar behavior threshold, confirm them as friend relationship, and output all friend pairs at the same time.
本发明优点是:能够利用用户产生的高频刷卡数据有效识别用户在现实生活中的朋友关系。The invention has the advantages of being able to effectively identify the friendship of the user in real life by using the high-frequency card swiping data generated by the user.
附图说明Description of drawings
图1是一卡通数据样例图。Figure 1 is a sample diagram of one-card data.
图2是数据清理图。Figure 2 is a data cleaning diagram.
图3是进入图书馆数据图。Figure 3 is a data map of entering the library.
图4是离开图书馆数据图。Figure 4 is a graph of leaving library data.
图5是图书馆数据排序结果图。Figure 5 is a graph of library data sorting results.
图6是相似行为节点对图。Figure 6 is a graph of similar behavior node pairs.
图7是相似行为矩阵图。Fig. 7 is a similar behavior matrix diagram.
图8是帕累托图。Figure 8 is a Pareto chart.
具体实施方式detailed description
下面结合说明书附图1-8及实施例对本发明进一步详细说明。The present invention will be further described in detail below in conjunction with accompanying drawings 1-8 of the specification and embodiments.
一种基于高频刷卡数据的朋友关系识别方法,主要包括数据预处理、相似行为矩阵的生成、相似行为阈值的确定和朋友关系的识别;A friend relationship identification method based on high-frequency card swiping data, mainly including data preprocessing, generation of similar behavior matrix, determination of similar behavior threshold and identification of friend relationship;
1)数据预处理;获取当前所有用户的高频刷卡数据,提取数据中有效信息,包括用户的ID,刷卡的地点,刷卡的时间,刷卡操作的类型,删除数据中所有的标点,仅保留数字信息,并对数据进行统一的命名,存入统一的数据仓库中;1) Data preprocessing; obtain high-frequency card swiping data of all current users, extract valid information in the data, including user ID, location of swiping card, time of swiping card, type of card swiping operation, delete all punctuation in the data, and only keep numbers information, and uniformly name the data, and store them in a unified data warehouse;
2)相似行为矩阵的生成;2) Generation of similar behavior matrix;
(a)对于数据仓库中的数据而言,当任意两个用户的刷卡记录的地点一致,同时刷卡时间间隔不超过t秒时,则称这两个用户之间存在一次相似行为;(a) For the data in the data warehouse, when any two users have the same card swiping records at the same location, and the time interval between swiping cards at the same time does not exceed t seconds, it is said that there is a similar behavior between the two users;
(b)设定时间区间T,提取数据仓库中在时间区间T范围内的数据;在时间区间T范围内,用户m和用户n的相似行为次数就构成了相似行为矩阵的一个元素am,n;相似行为矩阵为主对角元素均为0的对称矩阵;(b) Set the time interval T, and extract the data within the time interval T in the data warehouse; within the time interval T, the number of similar behaviors of user m and user n constitutes an element a m of the similar behavior matrix, n ; The similar behavior matrix is a symmetric matrix whose main and diagonal elements are all 0;
3)相似行为阈值的确定;提取相似行为矩阵对角线上方的所有元素,生成帕累托图,取累计百分比首次超过α,0<α<100%所对应的相似行为的次数作为相似行为阈值,α通常取90%;3) Determination of the similar behavior threshold; extract all elements above the diagonal of the similar behavior matrix, generate a Pareto diagram, and take the number of similar behaviors corresponding to the cumulative percentage exceeding α for the first time, 0<α<100% as the similar behavior threshold , α usually takes 90%;
4)朋友关系的识别;找出所有相似行为超过相似行为阈值的用户对,将他们确认为朋友关系,同时输出所有的朋友对。4) Recognition of friend relationship; find out all user pairs whose similar behavior exceeds the similar behavior threshold, confirm them as friend relationship, and output all friend pairs at the same time.
实施例Example
1、数据预处理1. Data preprocessing
(1)数据清理(1) Data cleaning
从学校教务处获取的一卡通数据如图1所示,每一行代表一个刷卡记录,分别为学生的id,学生的刷卡操作,学生刷卡的时间。以10条数据为例,对数据进行清理。数据清理的内容包括删除不完整的数据,删除每一条数据中刷卡操作记录中的引号、刷卡时间记录中的引号和斜线号,经处理后得到结果如图2。将处理后的数据保存到数据仓库中。The one-card data obtained from the school’s academic affairs office is shown in Figure 1. Each row represents a card swiping record, which is the student’s id, the student’s card swiping operation, and the time when the student swiped the card. Take 10 pieces of data as an example to clean up the data. The content of data cleaning includes deleting incomplete data, deleting the quotation marks in the card swiping operation records in each piece of data, and the quotation marks and slashes in the card swiping time records. After processing, the result is shown in Figure 2. Save the processed data to the data warehouse.
(2)数据分类(2) Data classification
在每一条数据中,不仅记录了刷卡的时间,还记录了刷卡者是进入还是离开图书馆。由于两个用户如果存在朋友关系,通常都是同时进入图书馆或者同时离开。因此,我们假定只有两个用户同时进入或离开才称作一次相似行为。基于此种假定,我们将进入和离开的数据进行分类,建立两个txt文档,分别命名为“library_in”和“library_out”。“library_in”和“Library_out”中的数据分别如图3和图4所示。In each piece of data, not only the time of swiping the card is recorded, but also whether the swiper enters or leaves the library. Because if there is a friend relationship between two users, they usually enter the library at the same time or leave at the same time. Therefore, we assume that only two users entering or leaving at the same time is called a similar action. Based on this assumption, we classify the incoming and outgoing data, and create two txt files named "library_in" and "library_out". The data in "library_in" and "Library_out" are shown in Figure 3 and Figure 4, respectively.
(3)数据整理(3) Data collation
以“library_in”为例,使用C++sort排序方法将文本中的数据按时间排序,得到结果如图5所示。Taking "library_in" as an example, use the C++sort sorting method to sort the data in the text by time, and the result is shown in Figure 5.
2、相似行为矩阵的建立2. Establishment of similar behavior matrix
我们假定当两名用户刷卡地点一致,并且刷卡时间间隔不超过30秒时,则称两者存在一次相似行为。下面具体阐述相似行为矩阵的生成方法:We assume that when two users swipe their cards at the same location and the time interval between swiping their cards does not exceed 30 seconds, it is said that there is a similar behavior between the two users. The following is a detailed description of the generation method of the similar behavior matrix:
首先建立一个空的队列,之后按行遍历文本。如果队列为空,将读取到的当前行的时间信息排入队列;如果队列中存在信息,则将该行时间信息与队首的时间信息进行比较。若时间间隔小于30秒,则视为一次相似性为,输出到“friend”文本中,并继续向下遍历队列;如果时间间隔都小于30秒,则将该行信息加入到队尾中;如果时间间隔大于30秒,则将队列中该点的信息替换为文本中该行的信息,继续向下遍历文本。统计出两个时间间隔不超过30秒的同学id,得到朋友对,朋友对的两个id按递增顺序输出,并且保留朋友对的年月日信息。按半月为时间单位生成矩阵,输出结果如图6所示。将“friend”文档中的数据按半月为时间段分成若干个小文本,将用户id按递增顺序对应生成如图7所示的相似行为矩阵。First create an empty queue, then iterate through the text line by line. If the queue is empty, put the read time information of the current row into the queue; if there is information in the queue, compare the time information of this row with the time information of the head of the queue. If the time interval is less than 30 seconds, it is regarded as a similarity, output to the "friend" text, and continue to traverse the queue downward; if the time interval is less than 30 seconds, add the line information to the end of the queue; if If the time interval is greater than 30 seconds, replace the information of this point in the queue with the information of this line in the text, and continue to traverse the text downwards. Count the IDs of two classmates with a time interval of no more than 30 seconds, get a friend pair, and output the two IDs of the friend pair in increasing order, and keep the year, month, and day information of the friend pair. The matrix is generated according to the time unit of half month, and the output result is shown in Figure 6. The data in the "friend" document is divided into several small texts according to the time period of half a month, and the user id is correspondingly generated in increasing order as shown in Figure 7. Similar behavior matrix.
3、相似行为阈值的确定3. Determination of similar behavior threshold
我们设定α的值为90%,获取相似行为矩阵对角线上方的数据,生成帕累托图,如图8所示。由图8可知,相似行为次数为1的,累积百分比为85.04%。,相似行为次数为1和2时的累积百分比为96.38%,此时已经超过了设定的90%的阈值,因此取2为相似行为阈值。We set the value of α to 90%, obtain the data above the diagonal of the similar behavior matrix, and generate a Pareto diagram, as shown in Figure 8. It can be seen from Figure 8 that the cumulative percentage is 85.04% when the number of similar behaviors is 1. , when the number of similar behaviors is 1 and 2, the cumulative percentage is 96.38%, which has exceeded the set threshold of 90%, so 2 is taken as the threshold of similar behavior.
4.朋友关系的识别4. Recognition of friendship
我们将相似行为次数为2的用户对认定为存在朋友关系,因此用户id为9733和9773的学生被识别为朋友关系。We regard the user pair whose number of similar behaviors is 2 as having a friend relationship, so the students whose user IDs are 9733 and 9773 are identified as a friend relationship.
Claims (5)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710226852.6A CN106991182A (en) | 2017-04-06 | 2017-04-06 | A kind of friends recognition methods based on high frequency brushing card data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710226852.6A CN106991182A (en) | 2017-04-06 | 2017-04-06 | A kind of friends recognition methods based on high frequency brushing card data |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106991182A true CN106991182A (en) | 2017-07-28 |
Family
ID=59415828
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710226852.6A Pending CN106991182A (en) | 2017-04-06 | 2017-04-06 | A kind of friends recognition methods based on high frequency brushing card data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106991182A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111258994A (en) * | 2020-01-09 | 2020-06-09 | 华中师范大学 | A method for constructing social network for students' campus card behavior data |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103778689A (en) * | 2012-10-17 | 2014-05-07 | 刘全 | Personal-user-orientated access global control method and system |
US8888009B1 (en) * | 2012-02-14 | 2014-11-18 | Dynamics Inc. | Systems and methods for extended stripe mechanisms for magnetic cards and devices |
CN105991682A (en) * | 2015-01-30 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Data sharing method and device |
-
2017
- 2017-04-06 CN CN201710226852.6A patent/CN106991182A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8888009B1 (en) * | 2012-02-14 | 2014-11-18 | Dynamics Inc. | Systems and methods for extended stripe mechanisms for magnetic cards and devices |
CN103778689A (en) * | 2012-10-17 | 2014-05-07 | 刘全 | Personal-user-orientated access global control method and system |
CN105991682A (en) * | 2015-01-30 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Data sharing method and device |
Non-Patent Citations (1)
Title |
---|
童睿: "基于一卡通消费数据的学生成绩预测和朋友关系网络检测研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111258994A (en) * | 2020-01-09 | 2020-06-09 | 华中师范大学 | A method for constructing social network for students' campus card behavior data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111309759B (en) | Intelligent matching platform for enterprise science and technology projects | |
Wang et al. | Who sets the agenda? The dynamic agenda setting of the wildlife issue on social media | |
US10846474B2 (en) | Methods, devices and systems for data augmentation to improve fraud detection | |
El Ballouli et al. | Cat: Credibility analysis of arabic content on twitter | |
CN109584094B (en) | Interpersonal path rapid positioning system, method and medium | |
Lloyd et al. | Newspapers vs. blogs: Who gets the scoop? | |
Galiani et al. | Incentives for replication in economics | |
CN110134842B (en) | Information matching method and device based on information map, storage medium and server | |
Sun et al. | Do Airbnb’s “Superhosts” deserve the badge? An empirical study from China | |
Andleeb et al. | Identification and classification of cybercrimes using text mining technique | |
Comaroff et al. | After labor | |
CN109948138A (en) | A kind of comment processing method and system | |
Welch et al. | Learning from personal longitudinal dialog data | |
CN110046648A (en) | The method and device of business classification is carried out based at least one business disaggregated model | |
Barna et al. | Factors of digitalization of the marketing activity of tourist enterprises of Ukraine in the conditions of global digitalization | |
CN110175289B (en) | Mixed recommendation method based on cosine similarity collaborative filtering | |
CN107967299A (en) | The hot word extraction method and system of a kind of facing agricultural public sentiment | |
Zhang et al. | Modeling social attention for stock analysis: An influence propagation perspective | |
CN106991182A (en) | A kind of friends recognition methods based on high frequency brushing card data | |
Ni et al. | Rethinking big data in a networked world | |
Bharadwaj et al. | Source based fake news classification using machine learning | |
Ko et al. | Paid review and paid writer detection | |
Wickramasekara | Diasporas and development: Perspectives on definitions and contributions | |
Sunil Datt | The information explosion: Trends in technology 2011 review | |
Swartz et al. | Diversity from emojis and keywords in social media |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170728 |