Artificial Intelligence in Electronic Commerce

These silders share my ideas about what AI can do in EC. Generally, there are three issues.

• Logistics Management
• Recommendation Systems
• Cross-Monitor Advertisement

This issue includes

• product tracing, management and distribution
• price comparison and formulation
• warehouse allocation and management
• retailer and customer evaluation

Many retailers start to focus on this issue, and some companies offer B2B services to tackle and alleviate this heavy loading.

• Search system relies on users' active actions, however, with recommendation system, users can passively receive information.

• Emails can generally be viewed as a recommendation system. Nonetheless it is imprecise and annoying.

• With users' recent behaviors and some techniques (collaborative filtering and SVM), we can recommend what they like in “a suitable time”.

What is so called “suitable time”?

• In the morning, we recommend breakfast and coffee.
• In the afternoon, of course, afternoon tea coupon.
• In the weekend, movie tickets and department stores sales.
• Before the time stuffs have been used up, recommend users to buy next ones.

Rignt thing, right people, and right time.

According to research, about 59% people in Asia have three or more monitors.

• Do users behind different monitors have diffent hehaviors? Yes.

For example, users can use cellphone, desktop, and mobile web to get on 104.com.tw

• Most desktop users do not get on 104 in work hours.
• Females and younger prefer cellphone.
• People type fewer keywords in cellphone and type different keywords in different device.
  
• People stay longer in desktop.

We know their preferences more, and we can recommend them precisely.

• DATA, PRECISE DATA, and SUFFICIENT DATA.
• Data scientist plays like a cooker, to cook the data as a delicacy.
• He or she can not do behyond data.

• Cross-monitor data?
• different data in different time?
• the user's or product's information?
• other relative data.

In sum, we have some problem to solve, have we had or can we collect neceassary data to help us?

• Data plus Machine Learning plus Visualization = Problem Solution and Communication
• Using data and AI in EC is a popular topics, how to use and play the data is data scientist's duty. But how to collect and manage them is the mission for manager and data scientist.