Briefly describe the three key elements that make up the...

Briefly describe the three key elements that make up the information processing component of the MDSS. The information processing component of the MDSS consists of three key elements. These elements are data input, data processing, and data output. Data input involves the collection and entry of relevant data into the MDSS. This can include various types of data, such as weather data, traffic data, or patient data. The accuracy and completeness of the data input is crucial for the effectiveness of the MDSS. Data processing is the next step in the information processing component. This involves the analysis and manipulation of the collected data to generate useful information. Various algorithms and models are used to process the data and derive meaningful insights. This step may involve complex calculations and statistical analysis. Finally, data output is the presentation of the processed information to the end user. This can be in the form of reports, visualizations, or recommendations. The output should be clear, concise, and easily understandable for the user to make informed decisions.

Transcript

00:01 Okay, i'm going to do a markov chain.

00:03 Markov chain describes a process where the probability of the future depends only on the present.

00:10 We don't look at past states to determine the probability markov chains are used all over the place and used for all sorts of things.

00:17 The example we're going to look at is the learning algorithm for artificial intelligence, and i'm going to do some analysis of the probability in that change.

00:27 The first step to move from the diagram into the analysis is to create the matrix.

00:33 So this is the probability matrix what i've done here.

00:36 And i put it in dash line so i can keep my rows and columns straight, so i don't accidentally read the wrong data.

00:42 But what i've done here is to take st one and put it in this first row.

00:47 There is if you look at the diagram, there's a 0.45 probability that state one will move back in on itself and go to st one again.

00:55 There's a 10.45 probability that it will move to st to there are no arrows that go to three or four or five or seven or nine.

01:03 But there is a point, oh, three probability that it goes to six and 60.7 that it goes to eight.

01:09 And so i worked my way across state, one state to state three.

01:13 It's very important to keep that orientation.

01:15 So in state two, there's only an arrow that goes to three and has a one probability state.

01:21 Three only has an arrow going to state for, and it's a one probability note that these numbers should add up to one over here because there is a one probability that something will happen.

01:33 Even if it stays where it is, something will happen, so these numbers ought to add up to one.

01:38 If they don't, that's good.

01:39 Check that you've mis entered something.

01:42 Another thing to notice is down here.

01:44 There's an identity matrix that the eighth state has a one probability of going to the eighth state, and the ninth state has a one probability of going to the ninth state and over on over on the diagram that looks like this, and there's a little one next to it.

02:02 Those are absorbing states.

02:04 That's where the process will eventually end up.

02:06 So you give it enough time and the process ends up in these absorbing states.

02:10 And once they get there, they stay there to make the math a little easier for us to handle, a little more useful than we will say that it loops back on itself, and that allows us to put these ones here.

02:23 So now that we've done that, we have identified are absorbing states.

02:29 We have a matrix ready to go, then we can move on and look at meantime to absorption.

02:34 In the meantime, to absorption is how long or more properly, how many steps will it take to get from a particular stage to an absorbing stage? now we could construct more matrices, not just with one step in the future, but a matrix that shows two steps or three steps or four steps or 100 steps into the future.

02:55 And we could look at cumulative probability and multiply by the number of steps.

02:59 It would be an enormous headache, and it would only give us an estimate.

03:03 But there is a theorem that allows us to determine exactly what those mean times to absorption are.

03:10 And that's this theorem.

03:12 That's m t t a.

03:13 And we're going to use that and determine exactly how long it's going to take us now.

03:19 Before i do that, it shows up.

03:20 It has this queue in the equation.

03:23 And the lovely thing is, that's just a chunk that i'm going to pull out of this matrix.

03:28 This matrix is in what we call the canonical form.

03:32 The canonical form has it arranged so that the absorbing states are down here in the bottom rose.

03:37 However many there are they would be in the bottom rose.

03:40 If they're not down there, if you don't have the identity matrix down here, you would need to rearrange the matrix.

03:46 But ours has already set up.

03:47 And that means that over here, this chunk that does not include the absorbing states that is called q.

03:57 We're going to use that right away.

03:59 This other matrix over here that shows the probability of moving to those absorbing states that's called r.

04:08 Then over here i have an identity matrix.

04:11 You should see that in the canonical form, and then over here i have zeros.

04:16 So that's what i'm going to use to determine the meantime, to absorption.

04:21 The first step is to pull out q.

04:24 And so i have q right here.

04:27 So that's my cue matrix...

Question

Please give Ace some feedback