One of the poor's semantic processing toolbox: semantic version Jaccard

If of industry of machine learning (ML) practitioners of division of class, standard lines not you use the algorithm name sounds cool, regardless of your hand waving a is the tide depth up to the depth of the layer of 1000 learning nuclear bombs or sounds dregs of swords and spears naive Bayes, if you don't have a lot of data, especially can run supervised learning algorithm of labeled training data, you are standard in the field of ML bottom slag slag male female or slag Niang gun. Coupled with the computational resources, if your company has thousands of GPU cluster server can be for your drive, plus ten trains skin of training data, that you may become the upstart ml, large data processing field of Ma. ...
Read(723) comment(4)

Deep learning and natural language processing five: from RNN to LSTM

This paper introduces the basic technical principles of RNN and LSTM and its application in natural language processing. ...
Read(4019) comment(1)

A community question answering system based on convolutional neural network (CNN)

Q & a community is regarded as a kind of mature Internet applications, abroad such as quora stackoverflow, domestic such as old-fashioned Baidu know, a new generation of known almost, are considered is the representative of social Q & a community. Q & a community is the essence of personal meat knowledge base, through a period of time accumulation, will accumulate a lot of knowledge in the way of existence. ...
Read(2440) comment(1)

Application of deep learning in natural language processing (Version 0.76)

Deep learning, natural language processing...
Read(1651) comment(0)

Figure database Pregel

Excerpt from the big data RI Zhi: Proceedings of the architecture and algorithm "chapter fourteen, the books in this directory, pregel is proposed by Google for large-scale distributed graph computing platform, specifically to solve the web link analysis and social data mining applications involving large-scale distributed graph gauge problem. ...
Read(3375) comment(0)

Offline mining calculation model of large data graph database

Switchblade copyright statement to the: any reprint, reprint, please indicate the source and author information.*/: Zhang Junlin excerpt since the big data RI Zhi: Proceedings of the architecture and algorithm "ten four chapters and books in this directory for off-line mining class diagram for calculation of, at present has been the emergence of more outstanding, excellent and practical system with its own characteristics, the typical such as pregel and giraph, HAMA, PowerG...
Read(2600) comment(1)

The MapReduce of large data graph database is used for graph calculation

Switchblade copyright statement to: can any reproduced, reproduced please please indicate the source and author information.*/ CopyMiddle: Zhang Junlin excerpt from the big data RI Zhi: Proceedings of the architecture and algorithm "chapter fourteen, books in this directory. Using MapReduce graph calculation using MapReduce framework for large scale map data for the calculation is relatively less research. The main...
Read(3471) comment(0)

Data slice of large data graph database

Excerpt from the big data RI Zhi: Proceedings of the architecture and algorithm "chapter fourteen, books in this directory for mass to be data mining, in a distributed computing environment, facing the first problem is how to data evenly assigned to a different server. For non graph data, this problem is often more intuitive to solve, because there is no independent association between records, so there is no special constraint on the data segmentation algorithm, as long as the machine load as much as possible. As a result of the strong coupling between data records, if the data is not reasonable, not only will cause the imbalance between the machine load, but also a large increase in the machine...
Read(3020) comment(0)

TAO database of large data graph database

Excerpt from "big data structure and algorithm of" RI Zhi Lu: the fourteen chapter 14.1.2 TAO map database Is currently the world's most famous social networking site Facebook, if from the perspective of data abstraction, Facebook's social graph includes not only the relationship between friends, including the relationship between people and entities and entities, each user, each page, each picture, each application, each place and each comment are can be as an independent entity, the user like a page is the establishment of a relationship between users and web pages...
Read(2347) comment(4)

"Big data" RI Zhi Lu: Architecture and algorithm directory

4 directory editor In the zeroth chapter we talk about what we are talking about when we talk about big data... What is the 0.1 big data... 2 0.2 wings: big data technology paradigm transformation....................................... 4. 0.3 big data business alchemy.............................. 6...
Read(2637) comment(0)

"Big data: the architecture and algorithm of Gu Yanwu"

"Big data: the architecture and algorithm of Gu Yanwu" As mobile Internet o2o, wearable devices, such as the concept, "big data from Fu a proposed to storm through and swept the world, from the initial technical terms to penetrate the formation of social phenomenon in all walks of life, the time spent only a few years only, its also Bo Xing Yan. So, big data would like many once hot now trail is the popular notion as. Someday, people calmly raised his eyes found the wind has to and torpid, Tuliu under the setting sun sparkling ripples, people can not help but lament the dying suddenly Yan? Background of the book ...
Read(2566) comment(4)

HipHop algorithm: the use of micro Bo interactive relationship mining social circle

* Copyright: can be reproduced, please indicate the source and author information.*/ reprint CopyMiddle: Zhang Junlin TimeStamp:2012 in March   In micro blog environment, how to automatically mining a twitter user's social circle or circle of interest is a fundamental and important problem. If you can...
Read(4515) comment(2)

Text summarization technology research

* Copyright: can be reproduced, please indicate the source and author information.*/ reprint Text summarization technology research                            CopyMiddle: Zhang Junlin TimeStamp:2010 September One. Text...
Read(4695) comment(0)

Search engine anti cheating: the overall technical ideas

This article is an excerpt from "this is the search engine: the core technology of the eighth chapter." Above the, current search engine cheating means varied, emerge in endlessly, as should be the other side of the search engine, have also been adjusted accordingly technical ideas, we have to put forward of the anti spam technology scheme, so if finishing anti cheating technology scheme, you will find a lot of technical methods, clear thinking is not easy. Nevertheless, if for most anti spam technology in-depth analysis, will be found in the thinking of the overall technology still have rules to follow. From the basic point of view, can be roughly divided into the following three kinds of anti cheating:...
Read(7590) comment(1)

Search engine anti cheat: link cheating and hidden cheating

This article is an excerpt from "this is the search engine: the core technology of the eighth chapter." 8.2 link cheating So-called "link spam", website owners consider to search engine rankings using the "link analysis" technology, so by manipulating links between pages, or between pages link anchor text manipulation, in order to increase link ranking factor score, and affect the search results ranking methods of cheating. Common link cheating method is numerous, this section briefly introduces several popular cheating methods.   1 link farm (Farm Link) To...
Read(3467) comment(0)

The content of anti cheating in search engine

This article is an excerpt from "this is the search engine: the core technology of the eighth chapter." Page anti cheating is currently all commercial search engines need to address an important and difficult, for commercial profit driven, many webmaster will for search engine rankings were analyzed, and take some measures to improve website ranking, itself of this kind of behavior is understandable, many optimization behavior is accord with the search engine ranking rules, but there are also some malicious optimization, by means of a special will be web search ranking to improve and web quality is not commensurate with the position, this will seriously affect the search engine users search experience.
Read(6722) comment(1)

How to update the index of search engine index

This article is an excerpt from "this is the search engine: the core technology of the third chapter." Dynamic index through the maintenance of a temporary index in memory, can realize the dynamic real-time search and document support. However, the server memory is always limited, with the new system to add more and more documents, the temporary storage of the index will also increase the memory. When initially allocated memory will be used to consider updating the index disk with the contents of the temporary index, in order to release the memory space to accommodate the new document follow-up, this time to consider reasonable and effective index update strategy. There are four kinds of commonly used index updating strategies: total weight...
Read(4737) comment(2)

How to build index for search engine index

This article is an excerpt from "this is the search engine: the core technology of the third chapter." 3.4 establish index As the preceding sections of the, index structure if the establishment of a good, you can increase the speed of search, then the given a document collection, indexing is how to build up? There are many ways to set up the index, this section describes the three ways to establish a more practical method. Two times 3.4.1 (2-Pass In-Memory Inversion document traversal method)   As the name suggests, this method need to document...
Read(14302) comment(3)

Index based on search engine index

This article is an excerpt from "this is the search engine: the core technology of the third chapter." This section by introducing a simple example, introduction and search engine index related to some basic concepts, the understanding of these basic concepts for further knowledge indexing mechanism is very important.   3.1.1 word document matrix Word document matrix expression between the two is a contains the conceptual model of the relationship between, Figure 3-1 shows the meaning. Figure 3-1 each column represents a document, one for each word, on behalf of the inclusion relations hit the mark position. ...
Read(7489) comment(3)

Search engine link algorithm: HITS algorithm analysis

This article is an excerpt from "this is the search engine: the core technology of the sixth chapter." HITS algorithm is also linked analysis in a very basic and important algorithm, has now been Teoma search engine (www.teoma.com) as the link analysis algorithm in practical use. Hub 6.4.1 page and Authority page   Hub pages and Authority pages HITS algorithm is the most basic definition of two. The so-called "Authority" page refers to a field or a certain...
Read(7674) comment(1)
80 data a total of 4 pagesOne Two Three Four Next page Shadowe
    personal data
    • Visit311528 times
    • Integral:Four thousand one hundred and thirty-eight
    • Grade
    • Rank:4075th name
    • Original78
    • Reproduced:1
    • Translation:1
    • Comments:154
    Latest comments
    Friendship link