"Text analytics" of Information intelligence

What can be done with data available in the Internet, in social nets, or in the enterprises?

• Send, send, publish, edit, archive.
• Look for and index
• Categorize and classify according to tasks of the enterprise
• Take information and analyze
• Check hypotheses and ^[1]

What is "the text analytics"?

THAT integral part of the solution ADVANCED BI

Main opportunities of TA

Getting of data

• Extraction of the different sources given from a set. Accent shift from search to processing of the taken data
• Search and information extraction in real time
• Adaptation of the connector to data source at change of a data structure. Automatic analysis of unstructured documents, selection of heading, main part of the document, analysis of structure of the document and links
• Determination of data types. The choice of the processor depending on data type. Processing of audio, and video of information
• Ample opportunities on extraction of data from social nets, public data sources, the most visited sites and specialized sources

Categorization

• Document clustering, creation and filling of categories in documents. Selection of paragraphs and sections of documents using clustering algorithms
• Support of Pull (autorubrication) and Push (the predetermined headings) of scenarios for categories
• Grouping and categorization of documents and parts of documents according to the set scenarios
• Use of complex algorithms for a categorization. Use of the self-trained (statistical) algorithms and algorithms of supervised learning
• Maintaining [different] categories for different roles of users
• Use of categories for the subsequent information extraction from the unstructured text

Information extraction

• Extraction of significant information from the text
• Selection of own names, people, the organizations, geographical places, names, links to the websites, e-mail, messages at forums, messages in social nets and so forth.
• Selection of products/services and their characteristics the, including characteristics connected with customer service
• Filling by data of CRM systems. Search and filling by data on profiles of clients, the relations between clients, identification of family relations, identification of extent of influence between clients
• Extraction of other information relevant for marketing (a brendtreking, positioning in comparison with competitors and so forth)

Analysis of opinions

• The analysis of opinions is used for additional extraction of the hidden information which is difficult for taking using classical methods
• The analysis of opinions includes identification of tonality of messages, determination of an emotional component of the expression, selection of judgments, dreams and intentions of the client
• Use of tonality for scenarios of active removal of a negative from client side
• Separation of the facts and judgments
• The analysis of desires and dreams of clients for new product development and new marketing strategies

Machine learning

• Use of mathematical algorithms for training of the computer in the analysis of text information
• Use of complex training methods: statistical techniques, neural networks, method of entropy, decision tree and so forth.
• Use of machine learning together with manual analysis for quality improvement of work of TA
• Development of the new algorithms of machine learning directed to "imitation" the person for automation of "dialog with the client"

Application of TA for Voice of Client

Scope

The accelerated satisfaction of requirements, increase in a customer loyalty

• The cut of client impressions allowing to define customer needs and possibilities of increase in customer satisfaction
• Determination of key positive and negative drivers
• Active control of the relations with the client in interaction with the company

Increase in effective management of a brand and reputation

• Operational identification of cases of situations, negative for the company, and reaction to them before clients share experience with the friends or information will get to mass media
• Reaction to a negative from client side at that moment when it is still possible to correct

Modern approach to product lifecycle management

• Collection of information about preferences of clients concerning products and their advantages
• Use of strong and weak features of the competing products
• Collecting and the profound analysis of the arising trends, their identification and use
• Increase in product lifecycle due to identification of opportunities of use of the products "No name", segments of clients and product lines
• Identification of reaction of clients to a product yield in real time

Increase in efficiency of sales and marketing

• Identification of opportunities of up sale and cross sale
• The personalized offer for the clients using the Internet channel
• Measurement of influence from actions and campaigns
• Measurement of effect of the change in price

Identification of the most resonant trends

Significant improvement of customer service

• Reduction of the "intermediaries" listening to the client, reduction of distortions, decrease in outflow
• Reduction of volume of internal communications, automation and increase in efficiency of processing of client requests
• Early identification of the repeating requests and development of the acceptable solution
• Quality improvement of the knowledge base for independent problem solving by the client

Increase in efficiency of planning and design

• Automation of problems of collecting, categorization and reporting under a feedback
• Reduction of inevitable errors from a manual categorization and processing of a feedback
• Feedback processing reduction in cost
• It is more than time and efforts to be spent for business development and increase in revenue, but not for feedback processing

See Also

• Yandex Data Factory
• Brand Analytics: A monitoring system and the analysis of messages about a brand - Brand Analytics
• Vesolv - VoC - Voice of Customer. Solution for the analysis of opinions of clients
• Big Data
• Predicative (predictive) analytics

Notes