Product introduction of Zhongke Click (Beijing) Technology Co., Ltd.

(a), "information collection system" system overview:

Information collection refers to the whole process of real-time collection, extraction, mining and processing of customized target data sources by using computer software technology, thus providing data input for various information service systems.

Military dog information collection expert is a powerful, simple and practical Internet information collection and monitoring software based on artificial intelligence.

(2), Internet information collection and mining:

It is required to collect and monitor specific target data sources or non-specific target data sources from the Internet, extract information in a structured way and save it as a local structured database, and then combine it with other modules according to business process requirements to import applications and serve the electronic industry platform.

Internet data collection and mining technology refers to the whole process of real-time information collection, extraction, mining and processing of customized target data sources by using computer software technology, thus providing data input for various information service systems and publishing and analyzing data according to business requirements.

(3), the Internet collection system flow chart

Step 1: Determine the acquisition task.

Step 2: For each acquisition task, we have multiple target data sources.

Step 3: Make different collection configurations for different target data sources to ensure that data can be collected. Step 4: Schedule the acquisition task, update synchronously with the target site, and acquire incrementally.

Step 5: Collect data results and complete the process from heterogeneous to isomorphic data.

Step 6: Publish the data to the application platform through the publishing server.

(4) Eight application fields of military dog "information acquisition system"

1, search engine and vertical search 2, integrated portal and industry portal

3, e-government and e-commerce 4, knowledge management and knowledge * * *

5, enterprise competitive intelligence system 6, business intelligence system BI

7. Information consultation and information appreciation. Information security and information monitoring

(5) Military dog "information acquisition system"-software function

(1), clean filtering, intelligent text extraction, and image-text association.

(2) There are abundant data export interfaces, which can export data into various mainstream relational data structures.

(3) The military dog "information acquisition system" is simple in configuration.

For news information collection, just enter the address of the target website or the address of a theme page, and the software will automatically learn the style of the website and extract the information of the website. There is no need to configure templates, and the style of the target website changes, and the software will learn automatically. It provides an easy-to-understand field configuration wizard for data acquisition software, and maintenance personnel can configure any information acquisition with a little training. For complex collection process, information can be automatically collected and monitored through card scripts.

(4) Military dog "information acquisition system" takes what you get, and what you take is what you see.

(5) Incremental acquisition and automatic update of military dog "information acquisition system"

Increase collection: for the first collection target website, the software supports complete collection; For collected sites, incremental collection is supported. Support for automatic update: automatically detect whether the website has been updated, without missing any important information.

(6) The results collected by the "information collection system" of military dogs are automatically copied.

We don't judge by simple rules, but by the similarity of content, which has high accuracy and won't miss the judgment because of a little change in title or content. Even if the topic is completely changed, the system will judge correctly.

(7) The military dog "information acquisition system" has built-in powerful information monitoring.

You can monitor the related information of any website on the Internet through a keyword. You can also set up a monitoring channel to monitor information containing keywords collected by any site. For the numerical field, you can set the information that the monitoring error monitoring value appears in a certain range. Information monitoring reaches the field level. You can set monitoring properties for any collection target website, and the monitoring period reaches seconds. The changed information can be collected locally in a short time, and powerful site management tools can centrally manage and operate all collected objects.

(8) Military dog "information acquisition system" supports multiple codes.

Supports the encoding of various website information, such as GBK, BIG5, UNICODE and UTF8, and the software will automatically convert it into GBK code for unified processing. The software will automatically identify the organizational structure and coding of the website. Form management: customize the form at will to facilitate the collection of different contents, such as using a separate form for collecting software and a picture form for collecting pictures.

(9) Military dog "information acquisition system" can import and export information at will.

Provides information import and export, and can be seamlessly connected with other software. For example, CRM OA software provides a powerful function of importing and exporting information records, and you can import and export any channels and records. Can be exported to Excel/Access, etc. , or directly export to the specified database. When used in combination with an information publishing server, information can be published anywhere.

(10), military dog "information acquisition system" supports template reading.

For any information type, the software will automatically create a reading template, which is convenient for you to read quickly; You can customize a beautiful reading template for any information form, and you can also set different reading templates for any channel.

(1 1), multi-page content reorganization of military dog's information acquisition system.

When the article from the target data source is displayed in the page of the target website, the system can reorganize it automatically. The software runs stably, the data acquisition speed is fast, and it takes up less system resources.

After many transformations, the bottom module of software acquisition runs stably, with fast acquisition speed and less system resources. Multi-threads can run concurrently, and will not occupy too much system resources. The acquisition speed is fast enough to reach the position in an instant. The software can completely realize 7*24 hours uninterrupted unattended information collection. More detailed functions need to be experienced in use.

(12), list of other features of military dog "information acquisition system":

1 supports multiple languages: Simplified Chinese, Traditional Chinese, English, Japanese, Korean and other languages.

2. Support multiple website types: including html and rss.

3. Support login and collect after verification.

4. The software supports website information collection that requires login and verification code, and the collection process is completely manual.

5. Support attachment collection

Including image attachment set, multimedia attachment set, audio and video attachment set, and automatic mapping and association between attachments and text.

6. Fully structured extraction extracts the unstructured data of web pages into specific structured information data.

Web page search takes web page as the minimum unit, web page block analysis based on vision takes web page block as the minimum unit, and vertical search takes structured data as the minimum unit. These data are then stored in a database for further processing, such as deduplication, classification, etc. Finally, word segmentation and indexing can meet the needs of users through search.

In the whole process, data is extracted from unstructured data into structured data, and returned to users in an unstructured and structured way after deep processing.

7. The data is kept locally, and the information can be viewed at any time. The collect information will be automatically saved in that local database, and you can look it up at any time.

8, multi-line layer, multi-task

9. Support massive data collection.

10, the software is practical, easy to use and powerful.

1 1, portable, extensible and customizable.

(6) Configuration requirements of military dog "information acquisition system"

Requirements: WindowsNT4/ Windows 2000 Server or newer operating system.

Requirements: Microsoft SQL Server 7/ 2000 or other ODBC interface.

Requirements: Intel Xeon CPU or above, RAM or above, and hard disk space of 200GB or above.

(7) Performance of "information collection system" for military dogs

L, support multi-thread acquisition.

2. Single data acquisition is above G level.

3. The synchronous update of data and data source is less than 10 second.

4. The synchronous release of data is less than 10 second. (1) product background

"the wind rises at the end of qingping." The formation and development of public opinion in public crisis events is a process of gradual progress or decrease from discussion to order along several levels. When a public crisis breaks out, it is like hitting water with a stone, which often causes widespread concern of the masses, making relevant information quickly transmitted in a short time, and the amount of information per unit time is very large. Some irrational comments, gossip or negative reports often arouse people's general sense of crisis to a certain extent, and even affect people's trust in the party and government. The netizens' concern and reaction to the incident shocked the local government departments, which were under great pressure from public opinion. For enterprises, the wanton dissemination of negative information and the lack of necessary risk early warning means will affect the brand and development of enterprises and even bring a devastating blow to enterprises. Therefore, timely monitoring, collecting and judging online public opinion is an important prerequisite for guiding crisis public opinion.

At present, online public opinion is becoming an important basis for government administrative departments or enterprises to make decisions. Therefore, under the new situation, how to collect online public opinion information as soon as possible, track the development of the situation, inform the relevant departments in time, and deal with it quickly after each emergency is an urgent problem for the government and relevant functional departments of enterprises.

How to get to know the major events "related to me" in the first place?

How can I accurately collect the public opinion information of "I need it most"?

How to monitor these public opinion information on the whole network, leaving no dead ends? Important information "does not leak"!

How to prevent "invisible" things from happening on the network? Always know what the internet is doing!

How to prevent harmful information from spreading and public opinion from getting out of control, and prevent it from forming a climate?

How to trace back to the dissemination way of key content on the Internet? Internet public opinion can be "checked"!

How to predict the future trend of these public opinion information?

How to effectively guide and actively resolve the crisis of network public opinion?

How to deal with public emergencies on the Internet?

How to fully grasp social conditions and public opinion?

How to push online public opinion briefings and special reports for relevant departments at higher levels?

Based on the core technology of independent intellectual property rights and independent research and development, Zhongke Click (Beijing) Technology Co., Ltd. timely launched the military dog network public opinion monitoring system through in-depth investigation of the actual needs of the government and enterprises, combined with Zhongke Click's profound understanding of Internet public opinion management business and years of practical experience, which has been widely used in many national government agencies (policy research office, foreign publicity office, online publicity office, government affairs office, network management office) and large enterprises. Through the mature network public opinion monitoring tools, combined with the perfect leadership system and working mechanism, we can properly handle the network public opinion of public crisis events. Comprehensively analyze the development trend of online public opinion, and provide decision-making reference and risk early warning based on online public opinion monitoring. While providing public opinion monitoring system products, Zhongke Click Company has rich business accumulation and implementation experience in public opinion monitoring. It is the glorious mission and task of Clickman to provide advanced public opinion monitoring systems and services for the government, industry authorities and enterprises.

(2) Core technology

The network public opinion monitoring system is an advanced and powerful application system developed by Zhongke Click Company, which provides network public opinion monitoring and decision-making reference for the government and enterprises. Widely used in public opinion monitoring, competitive intelligence, risk early warning and other fields. Its main functions and performance are as follows:

The core technologies of network public opinion monitoring system are Internet information collection technology, natural language intelligent processing technology (text mining technology), full-text retrieval technology and public opinion application technology.

1. 1 Internet information collection technology

1. 1. 1 powerful information collection function

Powerful information collection function is the guarantee of all other functions. For products with less hard acquisition technology, it is impossible to achieve effective public opinion monitoring effect. The data collection and data mining of military dogs ranks first in the whole industry, which provides a strong guarantee for the deep processing of information.

1. 1.2 supports the monitoring of various network operators.

Can monitor major search engines, news portals, BBS, blogs, message boards. ...

1. 1.3 yuan data search function.

Meta-search engine integrates search engines with different performances and styles, and develops some new query functions. Checking a meta search engine is equivalent to checking multiple independent search engines. When searching and collecting network information, meta-search can specify search conditions, which not only improves the pertinence of information collection, but also expands the breadth of collection scope, with twice the result with half the effort.

1. 1.4 has thousands of monitoring websites.

You can easily monitor thousands of websites without too much configuration.

1. 1.5 websites that can monitor various languages and codes.

There is no need to configure automatic recognition language and website coding.

1. 1.6 intelligent information extraction technology

Web content intelligent extraction technology can effectively extract effective information from web pages, distinguish information items such as titles and words from web pages, automatically merge multiple web pages with continuous content, and automatically extract information from online forums.

1. 1.7 structured acquisition technology

Structured information extraction and data storage are carried out when collecting unstructured web data to meet the needs of multidimensional information mining and statistics.

1. 1.8 All-weather uninterrupted monitoring.

It can be monitored regularly or around the clock. Minute-level acquisition and update can be realized in practical application.

1.2 natural language intelligent processing technology

Automatic word segmentation technology of 1.2. 1

The word segmentation technology based on dictionary, rules and statistics is adopted to effectively solve the problem of word segmentation ambiguity. The language model method based on probability analysis is used comprehensively, which makes the accuracy of word segmentation reach 99%, and it can be used for word segmentation according to different applications with high speed.

1.2.2 automatic keywords and automatic summarization technology

Based on the semantic analysis of the text, the accurate automatic keywords and automatic summarization are realized by considering the word frequency, part of speech and location information comprehensively. At the same time, reference parsing and other technologies are used to make the abstract more readable.

1.2.3 automatic classification technology

Automatic classification technology without manual intervention can effectively improve the processing efficiency of unstructured information. Text classification refers to the function of computer to classify texts according to their contents. Zhongke click automatic classification technology includes the following two classification methods:

Automatic Text Classification Based on Content

Rule-based text classification

1.2.4 automatic clustering technology

Automatic clustering technology is an automatic clustering technology based on similarity algorithm, which automatically classifies a large number of unclassified documents, classifies documents with similar contents into one category, and automatically generates keywords for them, which provides convenience for determining category names. It can be used to automatically generate public opinion topics, track major news events and so on.

1.2.5 similarity retrieval and duplicate checking technology

Text duplicate checking technology based on document "fingerprint" supports information duplicate checking of massive data.

Similarity retrieval is a technique that refers to finding other texts with similar contents in a text set for a given sample. In practical application, find out articles with almost the same public opinion information and realize the elimination of public opinion information; According to the similarity of the theme of the article, a special report and background analysis are formed.

1.3 intelligent retrieval technology

The full-text engine of the system combines the traditional full-text retrieval technology with the latest WEB search technology, which greatly improves the performance index of the retrieval engine. At the same time, combined with a variety of related technologies, it provides rich retrieval methods and intelligent retrieval methods such as synonyms.

(3), product function

Military dog network public opinion monitoring system is the most mature network public opinion monitoring system and network public opinion office system independently developed by our company. The military dog network public opinion monitoring system is a platform that comprehensively uses search engine technology, text processing technology, knowledge management methods, natural language processing and mobile phone short messages. Through the automatic acquisition, extraction, classification, clustering, topic monitoring and topic focusing of massive information on the Internet, the users' demand for network public opinion monitoring and topic tracking of hot events can be met!

The system is based on the urgent needs of network public opinion monitoring and management, and is tailored for government departments, especially government propaganda departments. The system integrates the core functions of public opinion monitoring, public opinion collection, public opinion intelligent analysis, public opinion processing, public opinion early warning, public opinion search, public opinion report auxiliary generation, public opinion short message automatic reminder and so on. Help customers fully grasp the dynamics of public opinion and correctly guide public opinion. It plays an auxiliary role in ensuring the correctness of public opinion orientation of China's Internet mass media, sharing worries for the government, and monitoring and managing online public opinion. Using the military dog network public opinion monitoring system, the propaganda department can effectively regulate the internet information and guide healthy and beneficial public opinion orientation. The system has played a decisive role in promoting the strengthening of Internet information supervision, organizing forces to carry out information sorting and in-depth analysis, responding to public emergencies on the Internet, and comprehensively grasping social conditions and public opinion.

The military dog network public opinion monitoring system adopts the system architecture combining B/S and C/S structure, and adopts advanced system architecture to realize the browser-based client or ordinary client and server mode.

Military dog network public opinion monitoring system has been widely used in government propaganda departments at all levels and large listed companies. It has become an indispensable and trustworthy system for customers to monitor online public opinion.

1, powerful information collection function

Powerful information collection and data mining functions are the guarantee of all other functions. For products with less hard acquisition technology, it is impossible to achieve effective public opinion monitoring effect. The data collection and data mining of military dogs ranks first in the whole industry, which provides a strong guarantee for the deep processing of information.

2. Support the monitoring of various network operators.

Can monitor major search engines, news portals, BBS, blogs, message boards. ...

3. Thousands of monitoring websites are built in.

You can easily monitor thousands of websites without much configuration. ...

4. Websites that can monitor various languages and codes.

You need to configure automatic recognition language and website coding. ...

5. Intelligently extract the text and title of information

There is no need to configure automatic analysis to eliminate useless code such as advertisements. ...

6, all-weather uninterrupted monitoring

It can be monitored regularly or around the clock. ...

7. Automatically obtain the popularity of public opinion information and generate a report.

In the form of words and charts, various charts of network public opinion trends are generated in an intuitive form.

8. Real-time acquisition and monitoring of consultation clicks and responses, tracking poster information, poster IP, etc.

According to the number of views, reply, track the poster information, poster IP and other functions can let you know the attention and information source.

9. Public opinion information can be managed, searched, exported, edited, marked and classified.

You can manage and edit information, classify and mark information that you think is important, and facilitate the analysis and handling of similar events.

10, public opinion information can be further filtered.

Filter out public opinion information that is of great importance and urgently needs to be processed, and filter out useless, outdated and low-impact information.

1 1. The monitoring results are saved as historical snapshots, and the keywords in the article can be displayed incrementally (discolored).

The incremental display of keywords allows you to find the monitoring keywords and analyze their specific content in the first time, and the historical snapshot allows the content to reappear.

12, rich data interface, which can connect monitoring data with various systems.

13, automatically obtain the proxy IP function, which can prevent individual websites from anti-collection and anti-monitoring.

Collecting a large amount of website information for a long time will attract the attention of the website and may lead to blocking your IP. This situation can be effectively solved by automatically obtaining the IP address of the agent and replacing it in time.

14, public opinion report

Public opinion reports can be generated by selecting and dragging from channel navigation, channel monitoring or search results to another folder. You can choose the public opinion template that comes with the system or customize the public opinion template when outputting the public opinion report. Public opinion reports are finally provided to users in the form of world documents or web pages.

15, thermal analysis

Analyze popularity through article reprints, clicks and replies. Relevant data is stored in the database, and whether the link is active or not is automatically checked.

16, reproduced and disseminated

Analyze the propagation path of network public opinion, and analyze the website name corresponding to URL through reverse parsing technology.

17, SMS interface

By customizing the hot spot discovery rules, public opinion hot spots can be discovered automatically in time. And inform public opinion monitoring personnel in time through SMS to help them keep abreast of public opinion trends.

18, public opinion collaborative office platform

By allocating the relevant authority of different public opinion monitoring officers, it is convenient for the monitors to perform their duties more efficiently and master the public opinion dynamics. The demand of enterprise retrieval

1, heterogeneous data integration

Enterprise users need to search data from Internet sites and internal sites. There are both web pages and various database forms; There are not only structured data, but also unstructured and semi-structured data in various electronic file formats, such as Word, Excel, Lotus Notes, PDF, XML and so on. There are both text data and multimedia data; In addition, the data of the same organization may also be distributed on different media carriers.

However, no matter how different the form, source, location and platform of data are, enterprise users always hope that internal and external data can be seamlessly combined, and all resources can be searched with a single search tool and a unified interface, and satisfactory results can be obtained soon. Moreover, the content of internet search is unknown to users, while the objects of enterprise-level search are basically known information sources, including enterprise databases, directories, file systems, application systems and so on. When indexing this information, users need to arrange it according to the content, rather than comparing the source links.

2. Strict security search

Many people in the industry are worried about the topic of search security. They generally believe that the search environment is not fully prepared for enterprise applications, and the future is full of too many variables. However, in some practical applications, we can see that even if the data is defined with double security guarantees at the document level and the database level, the claws of search engines can search it through authorized index documents.

Therefore, different users in the enterprise network may have different access rights to different resources, which requires enterprise search engines to manage and control users, resources and rights at different levels to ensure the security of the system.

3, high reliability, comprehensive and accurate inspection.

As professional users, enterprise users need to find information with strong professionalism and complex concepts, and have very high requirements for the recall and precision of queries. Therefore, it is necessary to use various means to improve the precision and recall of search engines.

From the perspective of recall rate, internet search engines can't talk about recall rate, because the information on the Internet is overwhelming, and it is impossible for any search engine service provider to exhaust every page on the Internet. However, in some enterprise applications, missing retrieval is not allowed. It is necessary to index every piece of information that needs to provide services in the enterprise. On the premise of ensuring the efficiency of the retrieval mechanism, it can meet the requirements of comprehensive retrieval.

In the same way, on the Internet, due to the characteristics of freedom of information, it is decided that search can only be achieved through the core retrieval means of "keyword matching". In an enterprise, the organization of information is much more complicated. Enterprise search engine has a perfect information classification system, multi-layer logical organization form of metadata and object data, which meets the requirements of accurate query and metadata indexing system based on object data content.

4. Intelligent retrieval service

Search services within enterprises have distinct commercial characteristics, unlike Internet search engines which only provide information reference. Search results within the enterprise will directly participate in the operation and decision-making of the enterprise. Therefore, for the processing of search results, it is very important to use relevant intelligent technologies in the search process to achieve rapid, accurate and comprehensive positioning of target information.

Enterprise search engines are usually organically combined with other IT applications in enterprises.

Supported by the framework of content management technology and search technology, enterprise search engines are usually closely integrated with data management, content management, record management, competitive intelligence, teamwork, process management, information portal and other aspects of knowledge management to form a complete and flexible system for managing enterprise knowledge assets.

5. Real-time information search service

The internal search service of an enterprise has business characteristics, so it is necessary to participate the search results in the business decision-making of the enterprise. Therefore, the services provided by search engines must be able to dynamically reflect the actual situation, that is, when internal information changes, they must be able to respond in real time.

Search scheme of military dog enterprise