11/09/58 1 Computer Networks and Applications Sunantha Sodsee Information Technology King Mongkut’s University of Technology North Bangkok

2 11/09/58 Computer Networks [6] Types of Networks  Each computer or user in a network is referred to as a node.  The interconnection between the nodes is referred to as the communication link.  In most networks, each node is a personal computer, but in some cases a peripheral device such as a printer can be a node.

3 11/09/58 Computer Networks [6] The number of links L required between N PCs (nodes) is determined by using the formula L = N(N−1) / 2

4 11/09/58 Network Fundamentals [6] A network of four PCs.

5 11/09/58 Network Fundamentals [6] A star LAN configuration with a server as the controlling computer.

6 11/09/58 Network Fundamentals [6] A ring LAN configuration.

7 11/09/58 Network Fundamentals [6] A bus LAN configuration.

8 11/09/58 Internet Applications [6] The Internet is a worldwide interconnection of computers by means of a complex network of many networks. Anyone can connect to the Internet for the purpose of communicating and sharing information with almost any other computer on the Internet. The Internet is a communication system that accomplishes one of three broad uses:  Share resources  Share files or data  Communication.

9 11/09/58 Internet Applications [6] The primary applications of the Internet are:   File transfer  The World Wide Web  E-commerce  Searches  Voice over Internet Protocol  Video

10 11/09/58 Internet Applications [6] is the exchange of notes, letters, memos, and other personal communication by way of e- mail software and service companies. File transfer refers to the ability to transfer files of data or software from one computer to another. The World Wide Web is a specialized part of the Internet where companies, organizations, the government, or individuals can post information for others to access and use.

11 11/09/58 Internet Applications [6] E-commerce refers to doing business over the Internet and other computer networks, usually buying and selling goods and services by way of the Web. An Internet search allows a person to look for information on any given topic. Several companies offer the use of free search “engines,” which are specialized software that can look for websites related to the desired search topic.

12 11/09/58 Internet Applications [6] Voice over Internet Protocol (VoIP) is the technique of replacing standard telephone service with a digital voice version with calls taking place over the Internet. Video over Internet Protocol.  Video or TV over the Internet (IPTV) is becoming more common. The video (and accompanying audio) is digitized, compressed, and sent via the Internet. It is expected to gradually replace some video transmitted over the air and by cable television systems.

13 11/09/58 World Wide Web A system of globally unique identifiers for resources on the Web and elsewhere,  the Universal Document Identifier (UDI),  later known as Uniform Resource Locator (URL) andUniform Resource Locator  Uniform Resource Identifier (URI); Uniform Resource Identifier The publishing language HyperText Markup Language (HTML);HyperText Markup Language The Hypertext Transfer Protocol (HTTP).Hypertext Transfer Protocol

14 11/09/58

15 11/09/58 World Wide Web

16 11/09/58 E-Commerce How to enhance E-commerce sales?  Browsers into buyers  Cross-sell Recommender Systems!!

17 11/09/58 What are recommender systems? Recommender systems are systems which provide recommendations to a user  Too much information (information overload)  Users have too many choices Recommend different products for users, suited to their tastes.  Assist users in finding information  Reduce search and navigation time

18 11/09/58 Case Study: Amazon

19 11/09/58

20 11/09/58

21 11/09/58

22 11/09/58 Personalized Product Recommendation?

23 11/09/58

24 11/09/58

26 11/09/58 Which Sources of Information? Sources of information for recommendations: [1] Browsing and searching data Purchase data Feedback provided by the users Textual comments Expert recommendations Rating

27 11/09/58 Type of Recommendations [2] Population-based  The most popular news articles, or searches, or downloads  Frequently add content  No user tracking needed.

28 11/09/58 Type of Recommendations [2] Item-to-item  Content-based  One item is recommended based on the user’s indication that they like another item. If you like Lord of the Rings, you’ll like Legend.

29 11/09/58 Type of Recommendations [2] Challenges with item-to-item:  Getting users to tell you what they like Financial and time reasons  Getting enough data to make “novel” predictions. What users really want are recommendations for things they’re not aware of.

30 11/09/58 Type of Recommendations [2] Item-to-item  Most effective when you have metadata that lets you automatically relate items.  Genre, actors, director, etc. Also best when decoupled from payment  Users should have an incentive to rate items truthfully.

31 11/09/58 Type of Recommendations [2] User-based  “Users who bought X like Y.”  Each user is represented by a vector indicating his ratings for each product.  Users with a small distance between each other are similar.  Find a similar user and recommend things they like that you haven’t rated.

32 11/09/58 Type of Recommendations [2] User-based  Advantages: Users don’t need to rate much. No info about products needed. Easy to implement  Disadvantages Pushes users “toward the middle” – products with more ratings carry more weight. How to deal with new products? Many products and few users -> lots of things don’t get recommended.

33 11/09/58 Type of Recommendations: General [1] Content-based Recommender System  Recommend items similar to those users preferred in the past  User profiling is the key  Items/content usually denoted by keywords  Matching “user preferences” with “item characteristics” … works for textual information  Vector Space Model widely used

34 11/09/58 Type of Recommendations: General [1]  Not all content is well represented by keywords, e.g. images  Items represented by same set of features are indistinguishable  Overspecialization: unrated items not shown  Users with thousands of purchases is a problem  New user: No history available  Shouldn’t show items that are too different, or too similar

35 11/09/58 Type of Recommendations: General [1] Collaborative Recommender System Memory-based collaborative filtering techniques Main problems: scalability and handling of new users Model-based collaborative filtering techniques High accuracy of prediction No need for searching the whole user-item rating matrix (grouping users into models)

36 11/09/58 Type of Recommendations: General [1] Collaborative Recommender System  Use other users recommendations (ratings) to judge item’s utility  Key is to find users/user groups whose interests match with the current user  Vector Space model widely used (directions of vectors are user specified ratings)  More users, more ratings: better results  Can account for items dissimilar to the ones seen in the past too  Example: Movielens.orgMovielens.org

37 11/09/58 Type of Recommendations: General [1]  Different users might use different scales. Possible solution: weighted ratings, i.e. deviations from average rating  Finding similar users/user groups isn’t very easy  New user: No preferences available  New item: No ratings available  Demographic filtering is required  Multi-criteria ratings is required

38 11/09/58 Type of Recommendations: Example[1] Cluster Models  Create clusters or groups  Put a customer into a category  Classification simplifies the task of user matching  More scalability and performance  Lesser accuracy than normal collaborative filtering method

39 11/09/58 Type of Recommendations: Example[1] Item to item collaboration (one that Amazon.com uses)  Compute similarity between item pairs  Combine the similar items into recommendation list  Vector corresponds to an item, and directions correspond to customers who have purchased them  “Similar items” table built offline  Example: Amazon.comAmazon.com

40 11/09/58 Type of Recommendations: Example[1] Knowledge based RS  Use knowledge of users and items  Conversational Interaction used to establish current user preferences  i.e. “more like this”, “less like that”, “none of those” …  No user profiles maintained, preferences drawn through manual interaction  Query by example … tweaking the source example to fetch results

41 11/09/58 How RS Work? Similarity Measurement [4]  For two data objects, X = (x1, x2,..., xn) and Y =(y1, y2,..., yn), the popular Minkowski distance is defined as  where n is the dimension number of the object and xi, yi are the values of the ith dimension of object X and Y respectively, and q is a positive integer. When q = 1, d is Manhattan distance; when q = 2, d is Euclidian distance

42 11/09/58 How RS Work? Similarity w u, v between two users u and v, or w i, j between two items i and j, is measured by computing the Pearson correlation [4] where the i ∈ I summations are over the items that both the users u and v have rated and is the average rating of the co-rated items of the u-th user

43 11/09/58 Example

44 11/09/58 Prediction and Recommendation Computation To make a prediction for the active user, a, on a certain item, i, we can take a weighted average of all the ratings on that item according to the following formula [4]

45 11/09/58 Example

46 11/09/58 Example

47 11/09/58 Example

48 11/09/58 Challenging: # Users and # Items Clustering Algorithms [5]

49 11/09/58 Complex Networks Recommender Systems and Social Web

50 11/09/58 Complex Networks Realistic networks are Complex Networks  Biological Network: How the brain work efficiently?  Propagation Network: How viruses propagate through the computer?  Competitor network: How rumors spread out the human society?  Communication Network: How information transmission exchanges on the Internet ?

51 11/09/58 Biotech Industry in USA

52 11/09/58 Complex Networks What is a complex network?  Observes any form of user behavior Web surfing logs s transactions Communication over Blogs Friend lists Purchase history on e- commerce sites Any other kinds action that demonstrates user intent  It creates large scale graph from all this behavior data

53 11/09/58 Recommender Systems and Social Web [3]

54 11/09/58 Recommender Systems and Social Web [3] Facebook only allows a bidirectional connection among users  if user A is connected to B then B is also connected to A Twitter users can follow without being followed  user A is linked to B, B is not linked to A.

55 11/09/58 Recommender Systems and Social Web [3]

56 11/09/58 Recommender Systems and Social Web [3]

57 11/09/58 Recommender Systems and Social Web [3] If a user visited certain exhibits and her/his Facebook page mentions she/he is a "Fan" of certain items, those would be saved for later matching against new visitors profiles. New visitors would be recommended exhibits that were viewed by people whom they most resemble based on the items they are "Fan". Find user profiles resembling current visitor's profile, extract tagged photos that are also related to museum's key terms, recommend exhibits relating to those.

58 11/09/58 References [1] Aalap Kohojkar, Yang Liu, Zhan Shi, “Recommender Systems”, March 31, [2] Maria Fasli, “Agent Technology for e-Commerce”, Commerce.htmhttp://cswww.essex.ac.uk/staff/mfasli/ATe- Commerce.htm [3] Amit Tiroshi, Tsvi Kuflik, Judy Kay and Bob Kummerfeld, “Recommender Systems and the Social Web”, International Workshop at UMAP2011 on Augmenting User Models with Real World Experiences to Enhance Personalization and Adaptation, July 15, [4] Xiaoyuan Su, Taghi M. Khoshgoftaar, “A Survey of Collaborative Filtering Techniques”, Advances in Artificial Intelligence, Vol. 2009, [5] Badrul M. Sarwar, George Karypis, Joseph Konstan, and John Riedl, “Recommender Systems for Large-scale E-Commerce: Scalable Neighborhood Formation Using Clustering”, The Fifth International Conference on Computer and Information Technology (ICCIT 2002), [6] Louis E. Frenzel, Jr., “Principles of Electronic Communication Systems”, The third edition, McGraw- Hill, 2008.