CN111259236A - Recommendation method for donation crowd funding field - Google Patents
Recommendation method for donation crowd funding field Download PDFInfo
- Publication number
- CN111259236A CN111259236A CN202010022643.1A CN202010022643A CN111259236A CN 111259236 A CN111259236 A CN 111259236A CN 202010022643 A CN202010022643 A CN 202010022643A CN 111259236 A CN111259236 A CN 111259236A
- Authority
- CN
- China
- Prior art keywords
- donation
- data
- user
- crowd funding
- crowd
- 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
Images
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
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- 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)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a recommendation method for the donation crowd funding field, which is characterized by comprising the following steps: comprises the following steps: step 1: acquiring a data source through a web crawler, and performing local storage; step 2: carrying out data preprocessing on the acquired data to form standard data; and step 3: associating user intention donation crowd funding data: constructing a donation crowd funding knowledge base by a natural language method; extracting information of the text content of the user based on an entity information extraction technology; secondly, storing the extracted association relation between the user information and the public service information into a knowledge base; finally, performing knowledge retrieval and reasoning on a knowledge base to complete analysis of the incidence relation of the two donation intentions; and 4, step 4: based on a knowledge base, a recommendation method for user donation crowd funding is constructed: calculating the intention donation degree between the intentional donors by using a collaborative filtering algorithm according to the established donation crowd-funding incidence relation, and establishing an intention donation degree matrixWherein Sim (V)p,Vq) Representing the degree of intentional donation between stars with intentional donation, Vp and Vn represent users and commodities, rvpAnd rvqRespectively a matrix of two users.
Description
Technical Field
The invention relates to a recommendation system method for crowd funding information based on network donation, which combines methods of data mining, natural language processing and a recommendation system.
Background
With the development of the internet, the data characteristics of the internet are increasingly highlighted, the data volume of the internet is increased abnormally and rapidly, the data types are various, the data quality is good and uneven, and the association relationship is complex. Meanwhile, another outstanding characteristic of the internet big data is that the value density is low, the big data contains a large amount of repeated, noise and garbage data, and a large amount of concurrent but meaningless association modes exist. The large scale and complex association relation of the data enable the traditional text analysis and mining technology to be increased rapidly on the space-time complexity of calculation; in addition, the rapid data growth rate and the huge amount of data make the traditional full-scale computation mode no longer applicable. The inherent characteristics of the complexity of big data of the internet make information extraction very difficult. The recommendation system at present recommends the content of interest for the user mainly according to the similar characteristics of the user and the conditions of browsing data and the like. The crowd funding information is various in types and wide in recommendation range, and the user does not have similar characteristics or browsing records, so that the problem that the user is difficult to accurately push the information when cold starting occurs is solved.
Disclosure of Invention
The invention aims to solve the cold start problem of donation items and the poor effect on the current recommendation result, provides a donation information recommendation method based on text mining, and achieves the purpose of recommending donation or crowd funding item information for users; according to the scheme, the crawler acquires data of the star user, and the data are preprocessed to obtain a standard text; then constructing intention donation crowd funding data association; and finally, recommending through a collaborative filtering algorithm.
The technical scheme of the invention is as follows: in order to solve the problems, the invention adopts the idea of a collaborative filtering algorithm to find data such as similar characteristics of users, donation records and the like, firstly obtains the data of the users through a crawler, then adopts a text mining technology to mine data aiming at donation and crowd funding, and carries out association. Constructing a knowledge base by a natural language processing method to retrieve and reason knowledge; and finally, calculating the similarity according to a collaborative filtering algorithm, and realizing the recommendation of the donation crowd funding information.
A recommendation method for the donation crowd funding field comprises the following steps: step 1: acquiring a data source through a web crawler, and performing local storage; step 2: carrying out data preprocessing on the acquired data to form standard data; and step 3: associating user intention donation crowd funding data: constructing a donation crowd funding knowledge base by a natural language method; extracting information of the text content of the user based on an entity information extraction technology; secondly, storing the extracted association relation between the user information and the public service information into a knowledge base; finally, performing knowledge retrieval and reasoning on a knowledge base to complete analysis of the incidence relation of the two donation intentions; and 4, step 4: based on a knowledge base, a recommendation method for user donation crowd funding is constructed: calculating the intention donation degree between the intentional donors by using a collaborative filtering algorithm according to the established donation crowd-funding incidence relation, and establishing an intention donation degree matrix
Where Sim (Vp, Vq) represents the extent of intentional donation between stars of intentional donations, Vp and Vn represent users and commodities, and rvp and rvq are matrices of two users, respectively.
The data source obtained in the step 1 is public service information of the user, and the public service information comprises public service names, donation crowd funding conditions, public service places, cooperation institutions and comments.
And 2, the data preprocessing in the step 2 comprises data cleaning, redundancy removal, denoising and filtering operations to obtain normalized text data.
The invention has the beneficial effects that: the method for recommending donation crowd funding based on text mining divides a project into four steps: data acquisition, data cleaning, donation candidate crowd funding association relation building and recommendation algorithm building. And finally, the donation crowd funding information is recommended to interested star users through effective recommendation, so that the interested degree of the recommended users can be greatly improved.
Drawings
FIG. 1 is a web crawler technology roadmap;
FIG. 2 is a diagram of data preprocessing;
fig. 3 is a diagram of construction of a donation crowd funding knowledge base.
Detailed Description
The method mainly processes the acquired data through text mining and natural language processing methods to construct a relational database, searches and infers in the relational database, and calculates through a collaborative filtering algorithm.
The invention is further illustrated below with reference to examples or prior art solutions:
the method comprises the steps of firstly executing step 1, obtaining a data source, and obtaining data through platforms such as a microblog and a public service website;
the acquired data is mainly dynamic data and has updated data, such as information of public welfare of stars, regular donation conditions of enterprises and the like;
the data is crawled by writing python language through a pysider platform and then stored through a MongoDB database.
After the data is acquired, step 2 is executed to preprocess the data, and the following methods are mainly adopted for processing:
(1) data cleaning, cleaning data by filling missing values, smoothing noise, identifying or deleting outliers and resolving inconsistencies;
(2) data integration, namely combining and uniformly storing data in a plurality of data sources to construct a data warehouse;
(3) data conversion, namely performing data smoothing, data aggregation, data generalization and data normalization processing on the data;
(4) data specification is carried out by a data cube aggregation method, so that the scale of a data set is greatly reduced while the integrity of original data is approached or maintained;
secondly, executing step 3, associating intention donation crowd funding data;
the method adopted for the acquired text data is a method of entity identification and relationship extraction;
the method mainly comprises the steps of extracting and associating public service events of the star users, wherein the public service events comprise star user names, public service types, public service names, public service places, cooperation mechanisms and the like;
aiming at the entity recognition method in the field, a method based on a named entity dictionary is adopted;
the method based on the naming dictionary is to find the most similar words or phrases in the text data to complete the entity recognition by adopting a complete character string matching or partial character string matching mode;
adopting a method based on forward maximum matching to identify the entity according to the text characteristics;
in the practical process, named entity recognition which does not exist in a dictionary occurs, and recognition is carried out by using other rule methods;
finally, the entities of public service events, names, public service types, public service names, public service places, cooperation mechanisms and the like are identified.
Extracting the semantic relation among the entities to obtain the incidence relation of the donation crowd-funding field;
for example, the ancient heaven music, the public welfare schools and the ancient heaven music charity funds are information relations extracted from the ancient heaven music, the ancient heaven music is a name entity, the public welfare schools are public welfare type entities, the ancient heaven music charity funds are public welfare organization entities, and the three establish an association relation of donation, namely the ancient heaven music donates to the public welfare schools through the ancient heaven music charity funds;
the entity relationship in the donation crowd funding field is clear and simple;
extracting the relation of the entity obtained before by the user through a relation extraction method based on the rule;
formulating an extraction method, wherein if A represents a name, B represents a public welfare type, C represents a public welfare name, D represents a public welfare organization, and formulating the following relations;
a participates in C, A belongs to D, A participates in C belongs to B, and the like;
when the relation appears in the context of the entity of the document sentence, the star user is determined to have a relation to a certain public welfare activity and a donation will exist.
And storing the relation in a database, and reasoning the acquired relation by an inductive reasoning method.
Finally, executing the step 4, and constructing a recommendation method for the intention donation crowd funding;
calculating mainly according to the incidence relation of the donation crowd funded in the step 3;
let us set U ═ U1,u2,...umI, U, m and V, V1,V2,...VnV | ═ n, and for each VnThe intention has a triple representing the attribute information of the location point: avi ═ ViN,Vix,...ViyIn which V isiNID, V, indicating intentionixExpress intention ViThe degree of association of (c).
Each star user ujEvaluation information can be issued to each intention, where rijRepresenting user ujTo donation intention VnThe evaluation and desirability of their donations are published. The collaborative filtering algorithm is based on the information matrix r and the user ujThe current donation information is recommended for the intended donation users with the historical evaluation information and willingness. The specific calculation formula is as follows:
vp and Vn represent user and commodity, the existing form is stored in a matrix mode, which is the limited quantity of matrix dimension in popular meaning, and the numerator of the formula is the user calculated by matrix calculation and cosine functionThe product of the number of users, denominator is the modulo calculation, rvp and rvq are the matrices for two users, respectively. According to the formula, the user V is a star userpWith star user VqInformation matrix r and users ujThe historical evaluation information, the will and the like, calculate the similarity value, and take the value range of [ -1,1 ] according to the cosine similarity]And taking the public service commodity which is calculated to be the value closest to 1 as the star user, and finally recommending the donation user whose public service item is willing.
Claims (3)
1. A recommendation method for the donation crowd funding field is characterized in that: comprises the following steps: step 1: acquiring a data source through a web crawler, and performing local storage; step 2: carrying out data preprocessing on the acquired data to form standard data; and step 3: associating user intention donation crowd funding data: constructing a donation crowd funding knowledge base by a natural language method; extracting information of the text content of the user based on an entity information extraction technology; secondly, storing the extracted association relation between the user information and the public service information into a knowledge base; finally, performing knowledge retrieval and reasoning on a knowledge base to complete analysis of the incidence relation of the two donation intentions; and 4, step 4: based on a knowledge base, a recommendation method for user donation crowd funding is constructed: calculating the intention donation degree between the intentional donors by using a collaborative filtering algorithm according to the established donation crowd-funding incidence relation, and establishing an intention donation degree matrixWherein Sim (V)p,Vq) Representing the degree of intentional donation between stars with intentional donation, Vp and Vn represent users and commodities, rvpAnd rvqRespectively a matrix of two users.
2. The recommendation method for the donation crowd funding field according to claim 1, wherein: the data source obtained in the step 1 is public service information of the user, and the public service information comprises public service names, donation crowd funding conditions, public service places, cooperation institutions and comments.
3. The recommendation method for the donation crowd funding field according to claim 1, wherein: and 2, the data preprocessing in the step 2 comprises data cleaning, redundancy removal, denoising and filtering operations to obtain normalized text data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010022643.1A CN111259236A (en) | 2020-01-09 | 2020-01-09 | Recommendation method for donation crowd funding field |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010022643.1A CN111259236A (en) | 2020-01-09 | 2020-01-09 | Recommendation method for donation crowd funding field |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111259236A true CN111259236A (en) | 2020-06-09 |
Family
ID=70950342
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010022643.1A Pending CN111259236A (en) | 2020-01-09 | 2020-01-09 | Recommendation method for donation crowd funding field |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111259236A (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103593417A (en) * | 2013-10-25 | 2014-02-19 | 安徽教育网络出版有限公司 | Collaborative filtering recommendation method based on association rule prediction |
US20140067659A1 (en) * | 2012-08-29 | 2014-03-06 | David G. Drake | System and method for charitable donation handling |
CN104317900A (en) * | 2014-10-24 | 2015-01-28 | 重庆邮电大学 | Multiattribute collaborative filtering recommendation method oriented to social network |
CN105320719A (en) * | 2015-01-16 | 2016-02-10 | 焦点科技股份有限公司 | Crowdfunding website project recommendation method based on project tag and graphical relationship |
US20160086250A1 (en) * | 2014-09-24 | 2016-03-24 | Wipro Limited | Systems and methods for providing product recommendations |
CN106022931A (en) * | 2016-04-19 | 2016-10-12 | 李麟飞 | Public welfare social contact recommending method |
CN107423343A (en) * | 2017-05-12 | 2017-12-01 | 中国地质大学(武汉) | A kind of library book based on mixing collaborative filtering recommends method and system |
CN110148043A (en) * | 2019-03-01 | 2019-08-20 | 安徽省优质采科技发展有限责任公司 | The bid and purchase information recommendation system and recommended method of knowledge based map |
-
2020
- 2020-01-09 CN CN202010022643.1A patent/CN111259236A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140067659A1 (en) * | 2012-08-29 | 2014-03-06 | David G. Drake | System and method for charitable donation handling |
CN103593417A (en) * | 2013-10-25 | 2014-02-19 | 安徽教育网络出版有限公司 | Collaborative filtering recommendation method based on association rule prediction |
US20160086250A1 (en) * | 2014-09-24 | 2016-03-24 | Wipro Limited | Systems and methods for providing product recommendations |
CN104317900A (en) * | 2014-10-24 | 2015-01-28 | 重庆邮电大学 | Multiattribute collaborative filtering recommendation method oriented to social network |
CN105320719A (en) * | 2015-01-16 | 2016-02-10 | 焦点科技股份有限公司 | Crowdfunding website project recommendation method based on project tag and graphical relationship |
CN106022931A (en) * | 2016-04-19 | 2016-10-12 | 李麟飞 | Public welfare social contact recommending method |
CN107423343A (en) * | 2017-05-12 | 2017-12-01 | 中国地质大学(武汉) | A kind of library book based on mixing collaborative filtering recommends method and system |
CN110148043A (en) * | 2019-03-01 | 2019-08-20 | 安徽省优质采科技发展有限责任公司 | The bid and purchase information recommendation system and recommended method of knowledge based map |
Non-Patent Citations (1)
Title |
---|
蔡圆媛著: "大数据环境下基于知识整合的语义计算技术与应用", pages: 211 - 215 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220121695A1 (en) | Knowledge graph-based case retrieval method, device and equipment, and storage medium | |
CN110910243B (en) | Property right transaction method based on reconfigurable big data knowledge map technology | |
WO2020108608A1 (en) | Search result processing method, device, terminal, electronic device, and storage medium | |
CN109815308B (en) | Method and device for determining intention recognition model and method and device for searching intention recognition | |
US8589398B2 (en) | Search clustering | |
US20180158078A1 (en) | Computer device and method for predicting market demand of commodities | |
CN106708929B (en) | Video program searching method and device | |
CN107357793A (en) | Information recommendation method and device | |
CN109033132A (en) | The method and device of text and the main body degree of correlation are calculated using knowledge mapping | |
CN115599899B (en) | Intelligent question-answering method, system, equipment and medium based on aircraft knowledge graph | |
CN114254201A (en) | Recommendation method for science and technology project review experts | |
CN115270738A (en) | Method and system for generating newspaper and computer storage medium | |
CN105205163A (en) | Incremental learning multi-level binary-classification method of scientific news | |
CN106886565A (en) | A kind of basic house type auto-polymerization method | |
CN110674313B (en) | Method for dynamically updating knowledge graph based on user log | |
CN110795613A (en) | Commodity searching method, device and system and electronic equipment | |
CN112784049B (en) | Text data-oriented online social platform multi-element knowledge acquisition method | |
CN112685440B (en) | Structural query information expression method for marking search semantic role | |
CN116629258B (en) | Structured analysis method and system for judicial document based on complex information item data | |
Patwardhan et al. | ViTag: Automatic video tagging using segmentation and conceptual inference | |
US11295078B2 (en) | Portfolio-based text analytics tool | |
CN111753151A (en) | Service recommendation method based on internet user behaviors | |
CN111259236A (en) | Recommendation method for donation crowd funding field | |
CN115309995A (en) | Scientific and technological resource pushing method and device based on demand text | |
CN114637846A (en) | Video data processing method, video data processing device, computer equipment and storage medium |
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 |