use SWI-prolog only figue out step 6 give the solution

% DO NOT USE THESE PROLOG FEATURES IN THIS ASSIGNMENT!! YOU

% WILL NOT RECEIVE CREDIT FOR A SOLUTION CONTAINING THESE:

% "assert", ";", "->".

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% What to turn in:

% 1. A paper listing of this file containing your program and a

% paper copy of SCREEN SHOTS showing the result of running

% your program. Turn this in in class. If you are on a

% 2-person team, only turn in one paper copy (with both

% names on it).

% 2. Submit a copy of your program on Canvas named hw2.txt.

% The digital copy will be used in case the grader needs to

% run your program to verify that it works, and to determine

% the submission time if late. Each student on a team should

% submit a duplicate copy on Canvas under his/her own name.

% 3. Submit a copy of your SCREEN SHOTS on Canvas as

% a .pdf, .doc, or .docx file. Turn in one document with all

% screen shots. Do not use a phone to photograph the screen.

% Use "print screen" (PrtSc) and paste into a document.

% Eact student on a team should submit a copy on Canvas

% under his/her own name.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Problem: Use Bratko's BN interpreter in fig15_11_mod on Canvas.

% Build a Bayesian network like the one described in

% HW3-description.docx. Then ask the user if the patient lives

% in a high pollution area, smokes, coughs, and has a positive

% x-ray. Then use the BN to compute the probability that the

% patient has asthma or cancer and tell the user as in this

% dialog:

% Does the patient live in high pollution area? (y n):

% Does the patient smoke? (y n):

% Does the patient cough? (y n):

% Is the chest x-ray positive? (y n):

% Probability of asthma:

% Probability of cancer:

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Edit this file - add your solution to the steps below.

% Make your solution readable using indentation and white

% space. Do not use a line length that will wrap when printed.

% To run this file, rename it with a .pl extension if your are

% using SWI Prolog, or the proper extension for whatever Prolog

% you are using - you may use any standard Prolog interpreter.

% Directives (must be at the top of your SWI Prolog program).

% Syntax may differ in other Prolog interpreters.

:- dynamic p/2.

:- dynamic p/3.

% Step 1: Define the graph using the syntax required by

% Bratko's BN interpreter. Example: if there is an arc

% in the BN from node a to node b (a is the parent of b), say:

% parent(a, b).

% Note: Use lower case letters for nodes!!

parent(hp,a).

parent(a,co).

parent(s,ca).

parent(ca,co).

parent(ca,x).

% Step 2: Define the prior probabilities

% in the syntax for Bratko's BN interpreter.

% Example: p(b, 0.001). % prior probability of b is 0.001

% Note: Use lower case "p" in step 2 and step 3!!

p(hp,0.30).

p(s,0.20).

% Step 3: Define the conditional probability tables

% in the syntax for Bratko's BN interpreter.

% Note: the 1st argument is a node, the 2nd argument

% is a list of nodes or negated nodes, and the 3rd

% argument is the conditional probablity.

% Example syntax:

% p(sensor, [burglary, not(lightning)], 0.9).

% Note: the above in mathematical notation is:

% P(sensor | burglary & -lightening) = 0.9

p(a,[hp],0.7).

p(a,[not(hp)],0.2).

p(ca,[s],0.4).

p(ca,[not(s)],0.1).

p(co,[a,ca],0.9).

p(co,[not(a),ca],0.4).

p(co,[a,not(ca)],0.7).

p(co,[not(a),not(ca)],0.2).

p(x,[ca],0.9).

p(x,[not(ca)],0.03).

% Step 4: Follow these instructions for running your HW3

% with my modification of Bratko's BN interpreter:

% 1. Copy fig15_11_mod.pl (my mod of Bratko's BN interpreter)

% from CSC529 Canvas to your computer.

% 2. Start Prolog on your HW3 program. Before running your

% program, tell Prolog to consult fig15_11_mod.pl

% (in SWI Prolog, you would use menu commands: File

% menu -> Consult -> fig15_11_mod.pl)

% Step 5: Take screen shots showing the results of

% running the BN interpreter on these queries in

% this order:

% ?- prob(a, [hp, not(s), co, not(x)], P).

% ?- prob(a, [not(hp), s, co, x], P).

% ?- prob(a, [hp, s, co, x], P).

% ?- prob(a, [not(hp), not(s), co, not(x)], P).

% ?- prob(ca, [hp, not(s), co, not(x)], P).

% ?- prob(ca, [not(hp), s, co, x], P).

% ?- prob(ca, [hp, s, co, x], P).

% ?- prob(ca, [not(hp), not(s), co, not(x)], P).

% Step 6: Implement the dialog with the user here to

% call the BN with the values given by the user. Use

% the same questions as in the sample dialog below.

% It is not necessary to check for bad user input,

% but it should work for any valid answers!

% Sample dialogue:

% :- go.

% Does the patient live in high pollution area? (y n):

% Does the patient smoke? (y n):

% Does the patient cough? (y n):

% Is the chest x-ray positive? (y n):

% Probability of asthma: ___ (output the probability)

% Probability of cancer: ___ (output the probability)

% Give screen shots of the dialog for these user inputs:

% Pollution Smoke Cough X-ray

% y y y y

% y n y n

% n y y y

% n n y n

go :-write( 'Does the patient live in high pollution area? (y n):'),

read(choice),nl,

write(' Does the patient smoke? (y n):'),

read(choice2),nl,

write('Does the patient cough? (y n):'),

read(choice3),nl,

write('Is the chest x-ray positive? (y n):'),

read(choice4),nl.

process()

%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Figure 15.11 An interpreter for belief networks.

% Changes: changed member-> mymember, delete->mydelete,

% added parentheses to "not".

% Reasoning in belief networks

% Belief network is represented by relations:

% parent( ParentNode, Node)

% p( Node, ParentStates, Prob)

% where Prob is conditional probability of Node given

% values of parent variables ParentStates, for example:

% p( alarm, [ burglary, not earthquake], 0.99)

% p( Node, Prob)

% probability of node without parents

% prob( Event, Condition, P):

% probability of Event, given Cond, is P;

% Event is a variable, its negation, or a list

% of simple events representing their conjunction

prob( [X | Xs], Cond, P) :- !, % Probability of conjunction

prob( X, Cond, Px),

prob( Xs, [X | Cond], PRest),

P is Px * PRest.

prob( [], _, 1) :- !. % Empty conjunction

prob( X, Cond, 1) :-

mymember( X, Cond), !. % Cond implies X

prob( X, Cond, 0) :-

mymember( not(X), Cond), !. % Cond implies X is false

prob( not(X), Cond, P) :- !, % Probability of negation

prob( X, Cond, P0),

P is 1 - P0.

% Use Bayes rule if condition involves a descendant of X

prob( X, Cond0, P) :-

mydelete( Y, Cond0, Cond),

predecessor( X, Y), !, % Y is a descendant of X

prob( X, Cond, Px),

prob( Y, [X | Cond], PyGivenX),

prob( Y, Cond, Py),

P is Px * PyGivenX / Py. % Assuming Py > 0

% Cases when condition does not involve a descendant

prob( X, Cond, P) :-

p( X, P), !. % X a root cause - its probability given

prob( X, Cond, P) :- !,

findall( (CONDi,Pi), p(X,CONDi,Pi), CPlist), % Conditions on parents

sum_probs( CPlist, Cond, P).

% sum_probs( CondsProbs, Cond, WeigthedSum)

% CondsProbs is a list of conditions and corresponding probabilities,

% WeightedSum is weighted sum of probabilities of Conds given Cond

sum_probs( [], _, 0).

sum_probs( [ (COND1,P1) | CondsProbs], COND, P) :-

prob( COND1, COND, PC1),

sum_probs( CondsProbs, COND, PRest),

P is P1 * PC1 + PRest.

predecessor( X, not(Y)) :- !, % Negated variable Y

predecessor( X, Y).

predecessor( X, Y) :-

parent( X, Y).

predecessor( X, Z) :-

parent( X, Y),

predecessor( Y, Z).

mymember( X, [X | _]).

mymember( X, [_ | L]) :-

mymember( X, L).

mydelete( X, [X | L], L).

mydelete( X, [Y | L], [Y | L2]) :-

mydelete( X, L, L2).