how to fuse expert
play

How to Fuse Expert Need to Consider the . . . Knowledge: Not Always - PowerPoint PPT Presentation

Need to Fuse . . . Examples Fusing Expert . . . Example How to Fuse Expert Need to Consider the . . . Knowledge: Not Always In General, How . . . How to Define Degree . . . And but a Fuzzy Resulting Definition of . . . Discussion


  1. Need to Fuse . . . Examples Fusing Expert . . . Example How to Fuse Expert Need to Consider the . . . Knowledge: Not Always In General, How . . . How to Define Degree . . . “And” but a Fuzzy Resulting Definition of . . . Discussion Combination of Home Page “And” and “Or” Title Page ◭◭ ◮◮ Christian Servin 1 , Olga Kosheleva 2 , and Vladik Kreinovich 3 1 Computer Science and Information Technology Systems Department ◭ ◮ El Paso Community College, 919 Hunter Page 1 of 28 El Paso, TX 79915, USA, cservin@gmail.com 2 , 3 Departments of 2 Teacher Education and 3 Computer Science Go Back University of Texas at El Paso, El Paso, Texas 79968, USA olgak@utep.edu, vladik@utep.edu Full Screen Close Quit

  2. Need to Fuse . . . Examples 1. Need to Fuse Knowledge of Different Experts Fusing Expert . . . • Expert estimates of different quantities are usually not Example very accurate. Need to Consider the . . . In General, How . . . • In situations when measurements are possible, they are How to Define Degree . . . more accurate than expert estimates. Resulting Definition of . . . • When we can perform measurements: Discussion Home Page – we can further increase the measurement accuracy – if we use several different measuring instruments Title Page and then combine (“fuse”) their results. ◭◭ ◮◮ • It is known that such combinations are usually more ◭ ◮ accurate than all original measurement results. Page 2 of 28 Go Back Full Screen Close Quit

  3. Need to Fuse . . . Examples 2. Need to Fuse Knowledge (cont-d) Fusing Expert . . . • In many situations, measurements are not realistically Example possible, so we have to rely on expert estimates only. Need to Consider the . . . In General, How . . . • In such situations: How to Define Degree . . . – we can increase the accuracy of the resulting esti- Resulting Definition of . . . mates the same way: Discussion – by combining (fusing) estimates of several experts. Home Page Title Page ◭◭ ◮◮ ◭ ◮ Page 3 of 28 Go Back Full Screen Close Quit

  4. Need to Fuse . . . Examples 3. Examples Fusing Expert . . . • To estimate the temperature, we can ask two experts. Example Need to Consider the . . . • Suppose that: In General, How . . . – one expert states that the temperature is between How to Define Degree . . . 22 and 25 degree C, and Resulting Definition of . . . – another expert states the temperature is in the low Discussion seventies, i.e., between 70 and 75 F; Home Page – this corresponds to between 21 and 24 C. Title Page • Then we can conclude that the actual temperature is ◭◭ ◮◮ larger than 22 C and smaller than 24 C – i.e., the actual ◭ ◮ temperature is between 22 and 24 C. Page 4 of 28 • If we only ask one expert, we get an interval of width 3 that contains the actual temperature. Go Back • But by fusing the opinions of the two experts, we get Full Screen a narrower interval [22 , 24] of width 2. Close Quit

  5. Need to Fuse . . . Examples 4. Examples (cont-d) Fusing Expert . . . • So, we have indeed increased the accuracy. Example Need to Consider the . . . • Fusion is also possible on a non-quantitative level. In General, How . . . • For example, we can ask experts whether the wind is How to Define Degree . . . weak, moderate, or strong. Resulting Definition of . . . • Suppose that: Discussion Home Page – one expert says that the wind is not weak, while Title Page – another expert says that the wind is not strong. ◭◭ ◮◮ • By combining the opinions of both experts, we can ◭ ◮ conclude that the wind is moderate. Page 5 of 28 • On the other hand, if we only to one of the experts, we would not be able to come to this conclusion. Go Back Full Screen Close Quit

  6. Need to Fuse . . . Examples 5. Fusing Expert Knowledge: Non-Fuzzy Case Fusing Expert . . . • Let us start with the case when expert estimates are Example crisp (non-fuzzy). Need to Consider the . . . In General, How . . . • So, for each possible value of the estimated quantity, How to Define Degree . . . the expert is: Resulting Definition of . . . – either absolutely sure that this value is possible Discussion – or is absolutely sure that the given value is not Home Page possible. Title Page • In this case, each expert estimate provides us with a ◭◭ ◮◮ set of possible values of the corresponding quantity. ◭ ◮ • In most practical cases, this set is an interval [ x, x ]. Page 6 of 28 Go Back Full Screen Close Quit

  7. Need to Fuse . . . Examples 6. Non-Fuzzy Case (cont-d) Fusing Expert . . . • In these terms, when we have estimates of two different Example experts, this means that: Need to Consider the . . . In General, How . . . – based on the opinions of the first expert, we form How to Define Degree . . . a set S 1 of possible values; Resulting Definition of . . . – also, based on the opinions of the second expert, Discussion we form a set S 2 of possible values. Home Page • In general, different experts take into account different Title Page aspects of the situation. ◭◭ ◮◮ • For example, the first expert may know the upper bound ◭ ◮ x on the corresponding quantity. Page 7 of 28 • In this case, the set S 1 consists of all the numbers which are smaller than or equal to x , i.e., S 1 = ( −∞ , x ]. Go Back • The second expert may know the lower bound x , in Full Screen which case S 2 = [ x, ∞ ). Close Quit

  8. Need to Fuse . . . Examples 7. Non-Fuzzy Case (cont-d) Fusing Expert . . . • A natural way to fuse the knowledge is to consider Example numbers which are possible according to both experts. Need to Consider the . . . In General, How . . . • In mathematical terms, we consider the intersection How to Define Degree . . . S 1 ∩ S 2 of the two sets S 1 and S 2 . Resulting Definition of . . . • A problem occurs when this intersection is empty, i.e., Discussion when the opinions of two experts are inconsistent. Home Page • This happens: experts are human and can thus make Title Page mistakes. ◭◭ ◮◮ • In this case, an extreme option is to say that: ◭ ◮ – since experts are not consistent with each other, Page 8 of 28 – this means that we do not trust what each of them Go Back says, Full Screen – so we can as well ignore both opinions; the result of fusion is then the whole real line. Close Quit

  9. Need to Fuse . . . Examples 8. Non-Fuzzy Case (cont-d) Fusing Expert . . . • A more reasonable option is: Example Need to Consider the . . . – to conclude that, yes, both experts cannot be true, In General, How . . . but How to Define Degree . . . – we cannot conclude that both are wrong. Resulting Definition of . . . • They are experts after all, so it is reasonable to assume Discussion that one of them is right. Home Page • In this case, the result of the fusion is the union S 1 ∪ S 2 Title Page of the two sets. ◭◭ ◮◮ • In other words, the fusion S 1 f S 2 of the sets S 1 and S 2 ◭ ◮ has the following form; Page 9 of 28 – if S 1 ∩ S 2 � = ∅ , then S 1 f S 2 = S 1 ∩ S 2 ; Go Back – otherwise, if S 1 ∩ S 2 = ∅ , then S 1 f S 2 = S 1 ∪ S 2 . Full Screen Close Quit

  10. Need to Fuse . . . Examples 9. Example Fusing Expert . . . • Suppose that: Example Need to Consider the . . . – one expert says that the temperature is between 22 In General, How . . . and 25, and How to Define Degree . . . – another one claims that it is between 18 and 21. Resulting Definition of . . . • In this case, the intersection of the corresponding in- Discussion tervals [22 , 25] and [18 , 21] is empty. Home Page • This means that the experts cannot be both right. Title Page ◭◭ ◮◮ • What we can conclude: ◭ ◮ – if we still believe that one of them is right Page 10 of 28 – is that the temperature is either between 22 and 25 or between 18 and 21. Go Back Full Screen Close Quit

  11. Need to Fuse . . . Examples 10. Need to Consider the Fuzzy Case Fusing Expert . . . • In practice, experts are rarely absolutely confident about Example their opinions. Need to Consider the . . . In General, How . . . • Usually, they are only confident to a certain degree. How to Define Degree . . . • As a result, to adequately describe expert knowledge, Resulting Definition of . . . we need to describe: Discussion Home Page – for each number x , – the degree to which, according to this expert, the Title Page number x is possible. ◭◭ ◮◮ • This is the fuzzy logic approach; in the computer: ◭ ◮ – “true” (= “absolutely certain”) is usually repre- Page 11 of 28 sented as 1, and Go Back – “false” (= “absolutely certain this is false”) is rep- Full Screen resented as 0. Close Quit

Recommend


More recommend