Your secret is no longer a secret!
The book you want, Amazon understands you!
Facebook knows your preferences best!
You are responsible for being yourself, and Linkedin is responsible for guessing the TA you may be familiar with!
…………………………
And all this is inseparable.
big data
You still think
Big data = big data?
Tucson fault model
(Too young and too young)
What exactly is big data?
Come on! Weikeyun takes you to unveil the mystery of big data!
Data: It can be processed by computer.
Numbers: by manual processing
Watch a popular science video first ~
/iframe/preview.html？ vid = p 03026 a 40 u 9 & amp； Width = 500 & Height = 375 & Auto =0
(It is recommended to watch with WiFi)
① ① Large volume.
The amount of data is large! How big is it? Equivalent to n cloud disks with the size of 1T! ( 1T= 1024G！ ) 1T How old is it? 1G traffic always knows!
② There are many varieties.
Covers text, pictures, videos, logs, etc. Only you can't think of it, and you can't find it without it!
3 value is very valuable.
Don't underestimate these messy data! Their commercial value can help enterprises make profits ~!
4 speed is fast.
The processing speed is fast. At present, big data can handle more and more data, so more and more people eat soil after unemployment ~!
① All data are not random samples.
What matters is not the amount of data, but the greater the randomness! The higher the accuracy of the survey! It took the United States 13 years to work out the results of the early census. With big data, is your population 1 billion or1.300 million? Immediate results!
② The heterozygosity is inaccurate.
For example, Google, with its powerful translation function, collects messy data, so don't worry even if you don't know English ~!
③ Correlation is not causality.
Example: the relationship between income and happiness
-Causality: When income
-relevance: when income >; Twenty thousand dollars, the amount of money has nothing to do with happiness.
Set the timeline in the 20th century. At that time, when SRAS virus swept through Spain, it took at least half a month to be diagnosed. Half a month is not a long time, but it's a long time for the diagnosed patients ~ And the prediction data released by Google in the early days is nearly 97% consistent with the prediction data of CDC two months later, and the power of big data is amazing!
1 Li Yanhong's big data opens the intelligent era.
Speaking of big data, Li Yanhong, founder, chairman and CEO of Baidu, said in his speech "Big Data Opens the Intelligent Age", "Why is artificial intelligence so hot in recent years? One of the main reasons is because of big data ... "
In the New Year speech on 20 16, referring to artificial intelligence, Luo Pang said: The really most important battlefield has shifted to big data.
Baidu successfully used big data to "detect the heat map of scenic spots"
The heat index reflects the popularity of scenic spots and the ranking of urban heat.
Baidu map applies heat map to daily travel, calculates the crowd density and speed of clusters in scenic spots through location clustering, and comprehensively calculates the heat of cluster locations, thus reflecting the results in the "heat index".
Definition of clustering
The process of dividing a group of physical or abstract objects into multiple categories composed of similar objects is called clustering.
Liu Da Data and E-commerce
Talking about the value that the combination of big data and e-commerce can create for society, Liu believes that big data will improve the efficiency of the retail industry. JD.COM successfully tried to use big data analysis to improve logistics efficiency.
② Ma Da Data Ecology
Tencent has a clear positioning, and Ma Hua Teng hopes to do only one thing with his partners in the future:
Connect! Connect! Connect!
I hope to work with you to build the entire big data ecosystem.
Witkey Cloud applies big data to call centers.
With weiyun, BOSS and customer service staff, there is no need to worry about customer complaints, time-consuming report generation and data security. ...
Various applications based on speech recognition, combined with the application of big data analysis, have become the focus of attention of finance, insurance, government and major enterprises.
Using speech to text, through intelligent search and analysis, we can detect the keywords in the call, classify and cluster the content, and provide services such as call reason analysis, visual total quality inspection, telemarketing business analysis, and mining customer needs.
Micro-cloud focuses on the application import of voice big data analysis and related software development, and combines the most advanced recognition engine and analysis platform to launch a set of solutions. It is one of the technical leaders in the industry.
The basic techniques of speech recognition include feature extraction, comparison and matching and model optimization.
In practical application, speech recognition analysis is mainly based on speech-to-text (STT), that is, continuous speech analysis with large vocabulary. It is a pointer that recognizes continuous voice input and stores it in the writing font through voice inversion.
In the application of phonetic analysis of big dictionaries, it is necessary to establish a specific big dictionary according to the common words and expressions of users. Because this application is thesaurus comparison, different industries need to use different comparative thesaurus. One of the great advantages of Weikeyun is that it has its own professional team and complete analysis equipment and tools, which can modify and adjust the comparative thesaurus at any time according to the needs of customers and the changes of products and services to ensure the appropriate recognition rate.
Quality inspection, training and on-site management are the three pillars of call center operation. It has become the core foundation of management to find problems in quality inspection system and then emphasize improvement in training to improve the efficiency of operation management. As the most advanced quality inspection system in the industry, speech recognition quality inspection has the following features:
-Analyze all recordings from speech to text, and display the recording results in text form to improve the quality inspection efficiency.
-Describe the quality inspection rules in a logical way, conduct positive quality inspection, and analyze whether the customer service personnel introduce, verify and explain the relevant terms according to the service requirements.
-Logical way, testing service attitude, carrying out negative quality inspection, and analyzing whether the customer service staff are arrogant, rhetorical and impatient, so as to improve service quality.
-Give priority to major incidents according to the degree and time of violation.
-Flexible presentation of reports according to different quality inspection rules, organizational structure, degree of violation and occurrence period, so as to realize refined management.
Improvement and application of speech recognition pin
The main application of voice analysis in the field of bank telemarketing is credit card installment business, which is divided into three parts:
-Speech decomposition, which divides the sales process into five themes: building relationships, promoting demand, product description, objection handling, and transaction confirmation. Describe the communication content in a logical way, cluster, standardize the communication content and time allocation of agents, and standardize the speech.
-Customer label, which classifies customers according to communication content, such as questioning, caution, actual demand, conservatism, etc. According to the user's label, communicate with the person he likes.
-Loss analysis in the sales stage, knowing the reasons and proportion of potential users' rejection in each stage, making targeted analysis, designing retention tactics and improving the success rate.
-Invisibly supervise the seat to work hard; The daily conversation content of the agent will appear in the classified report the next day.
-Improve sales performance and increase the company's income.
-increase the confidence of agents, increase their income and reduce the turnover rate.
-Analyze the reasons for rejection and decide the next contact strategy, such as WeChat or EDM contact instead of direct reply.
-As a part of the source of big data analysis of banks, learn about consumption trends, product positioning and other information.
……
…………
For more details, please scan the code consulting assistant or search (micro signal: Wellcloud0 1).