monib Posted February 5, 2011 Report Share Posted February 5, 2011 Answer to your 1st qn:If you just add a trendline using your graphing software, it means you are using technology to find the function. In the first part you need to model a function analytically, which means you can't just click this and that and get a function in a few clicks! So yes you need to perform some calculations, which you did (i.e. using matrices). So stick with the matrices method.But in the 2nd part, you need to use technology to find a logistic function (researcher's model) which means you can just show the trendline using your graphing software to find the function! And in fact you must Answer to your 2nd qn:You only have 2 functions: the one you modelled and the researcher's model. The original data from the first table are just data points and not a model function! So now, you need to:1. Plot: data from table 1, your modelled function, data from IMF table2. Plot: data from table 1, the researcher's model, data from IMF tableSEPARATELY! and comment how well each model fits the IMF data Then after that pick one of those 2 models to be modified for the last part Thank you so much! Ok then i stick to matrices, great =) and Yeah I used my TI-84 Plus calculator to get the values of logistic funtion (reseracher's model), K, L and M (which means use of technology and i also described how i did it) Ok so i will only use the 2 functions then great (less work to do) hehe, but if i may ask (cause i didnt get it) i already have plotted the graphs...I started with the IMF data first and plotted a graph from the values then seperately I plotted first the modelled function (showing the table as well of the values) and then secondly the researcher's model so i have 3 graphs below each other kind of it you understand what i mean my question is what you meant exacly by "1. Plot: data from table 1, your modelled function, data from IMF table, 2. Plot: data from table 1, the researcher's model, data from IMF table" ?! should I plot the graphs (modelled function with IMF) and (researcher's model with IMF) together as well or is it enough if i have them seperately after each other ? Or do they mean thaty I SHOULD use the IMF data (those values) but with the other 2 models? insert the values from the IMF data into the 2 models i have?! =S i just read the question again and perhaps it should be like this instead?! Sry for all the questions Reply Link to post Share on other sites More sharing options...
dessskris Posted February 5, 2011 Report Share Posted February 5, 2011 Thank you so much! Ok then i stick to matrices, great =) and Yeah I used my TI-84 Plus calculator to get the values of logistic funtion (reseracher's model), K, L and M (which means use of technology and i also described how i did it) Ok so i will only use the 2 functions then great (less work to do) hehe, but if i may ask (cause i didnt get it) i already have plotted the graphs...I started with the IMF data first and plotted a graph from the values then seperately I plotted first the modelled function (showing the table as well of the values) and then secondly the researcher's model so i have 3 graphs below each other kind of it you understand what i mean my question is what you meant exacly by "1. Plot: data from table 1, your modelled function, data from IMF table, 2. Plot: data from table 1, the researcher's model, data from IMF table" ?! should I plot the graphs (modelled function with IMF) and (researcher's model with IMF) together as well or is it enough if i have them seperately after each other ? You are welcome! Ok that's great! Don't forget to put a screenshot, in case if you missed that!Okok I get that.. Hmm sorry that was rather confusing firstly you plot on the same axes the data from the first table together with the IMF data together with your modelled function and comment on how well it fits the IMF data. Secondly you plot on a new set of axes the data from the first table together with the IMF data together with the researcher's model and comment on how well it fits the IMF data. So you have them separately after each other. Well, you may also want to put everything on the same set of axes so that it would be easier to compare which one fits the data better but I somehow prefer to see separate graphs.Any further doubts or questions just ask again!Or do they mean thaty I SHOULD use the IMF data (those values) but with the other 2 models? insert the values from the IMF data into the 2 models i have?! =S i just read the question again and perhaps it should be like this instead?! Sry for all the questions I dont get what you mean what are you thinking of? lol I dont think so though, how can you insert the values from IMF data to the 2 models? you mean the values of x? then the values of y would be different from the ones given in the data :/and no problem, you better be asking than keeping quiet there is this indonesian proverb saying 'malu bertanya sesat di jalan' which means if you are afraid (or too shy) to ask you'll be lost Reply Link to post Share on other sites More sharing options...
monib Posted February 5, 2011 Report Share Posted February 5, 2011 So my last question: Well i described instead how i used the calculator but i think it will be enough!"firstly you plot on the same axes the data from the first table together with the IMF data together with your modelled function and comment on how well it fits the IMF data." this is what you wrote Which one is the data from the "FIRST TABLE" ?! = original data perhaps, the values from there?! Then I will have 3 curves on each graph, am i thinking right? Reply Link to post Share on other sites More sharing options...
dessskris Posted February 5, 2011 Report Share Posted February 5, 2011 So my last question: Well i described instead how i used the calculator but i think it will be enough!"firstly you plot on the same axes the data from the first table together with the IMF data together with your modelled function and comment on how well it fits the IMF data." this is what you wrote Which one is the data from the "FIRST TABLE" ?! = original data perhaps, the values from there?! Then I will have 3 curves on each graph, am i thinking right? Why do you have to have a curve for your data from the table? I meant the first table, the table on the top on the task sheet the original data yesssh. and I would actually just put the data points (from both the original data and the IMF data) there without connecting the dots and then put a model, so you only have 1 curve on 1 graph If you have any further question just ask here! I'm off to bed though so will probably answer them tomorrow. sorry! Reply Link to post Share on other sites More sharing options...
hannahlou Posted February 5, 2011 Report Share Posted February 5, 2011 HiThis is due in 2 days and I've nearly finished apart from one little bit so I would be really grateful if anyone could help!I've found values for K, L and M but I don't know how to interpret K, L and M - what do they mean???????Any help appreciated! Reply Link to post Share on other sites More sharing options...
fudger562 Posted February 5, 2011 Report Share Posted February 5, 2011 From what I'm reading it seems like statistics and matrices are a required for this portfolio but I haven't learned either yet.My first step to trying to figure out what type of function this is was to use finite differences to see if the function in the data or any of its derivatives increase at a constant pace. What I found was that it was increasing upwards on f(x) but every derivative afterward would randomly increase or decrease. Additionally, neither population or year can be 0 so that further proves my conclusions that its either a logarithmic or an exponential function. But now I have no clue where to go next with this portfolio. Can someone give me a general direction to go? Reply Link to post Share on other sites More sharing options...
greenglob Posted February 6, 2011 Report Share Posted February 6, 2011 Hi guys,Now this isn't one of those "My portfolio is due in two days and I need help and I just ask a bunch of questions looking for answers". I pretty much have everything set except for this one little part. I know that the "perfect" model for this Modelling portfolio would be the Logistic curve but I was wondering if I can determine the equation of the curve algebraically. Of course, I can simply use my GDC for a Logistic regression but I want to know if there is any way on how to come up with the Logistic equation algebraically. Thanks a bunch guys!If anyone wants help on this portfolio, I'll be glad to do as much as possible (without me just blatantly giving out the answer).-Greenglob Reply Link to post Share on other sites More sharing options...
dessskris Posted February 6, 2011 Report Share Posted February 6, 2011 HiThis is due in 2 days and I've nearly finished apart from one little bit so I would be really grateful if anyone could help!I've found values for K, L and M but I don't know how to interpret K, L and M - what do they mean???????Any help appreciated! Try to find what factors can affect population. Maybe ask a friend who takes Geography or a Geog teacher! From what I'm reading it seems like statistics and matrices are a required for this portfolio but I haven't learned either yet.My first step to trying to figure out what type of function this is was to use finite differences to see if the function in the data or any of its derivatives increase at a constant pace. What I found was that it was increasing upwards on f(x) but every derivative afterward would randomly increase or decrease. Additionally, neither population or year can be 0 so that further proves my conclusions that its either a logarithmic or an exponential function. But now I have no clue where to go next with this portfolio. Can someone give me a general direction to go?Statistics: not really, you basically just need to know how to use your GDC and/or your graphing softwareMatrices: it's just one way to solve the fastest way I should say, but elimination and substitution will doWow I never thought you'd need to do that hmm probably because I'm not doing this portfolio task. So you decide then: logarithmic or exponential. I don't quite understand why you need to deal with the difference&derivative, though, but if you are sure with that, go for it and just pick one type of function and start modelling! Start with stating the general formula for that type of function and then perform some calculations to solve for the constants and at the end you'll get the function!If still unclear just ask again!Hi guys,Now this isn't one of those "My portfolio is due in two days and I need help and I just ask a bunch of questions looking for answers". I pretty much have everything set except for this one little part. I know that the "perfect" model for this Modelling portfolio would be the Logistic curve but I was wondering if I can determine the equation of the curve algebraically. Of course, I can simply use my GDC for a Logistic regression but I want to know if there is any way on how to come up with the Logistic equation algebraically. Thanks a bunch guys!If anyone wants help on this portfolio, I'll be glad to do as much as possible (without me just blatantly giving out the answer).-GreenglobAre you sure you want to know? Of course there is! But I doubt you are expected to go to that extent. Besides, it says use technology so you don't solve algebraically.First, state the general formula.P = K / (1 + Le-Mt)I think the best and the simplest (and the only..? idk) way to solve algebraically is to substitute 3 sets of values of t and P there and then solve for K, L and M.Do you get that? I can give you an example if you want! Btw would you please not give help through PM or other methods? If you get any PM would you please ask them to post the questions here so if they can enjoy the 'hints' that you give, everybody can too well, just like what I've been doing here! Thank you! Reply Link to post Share on other sites More sharing options...
You Kyum Kim Posted February 6, 2011 Report Share Posted February 6, 2011 hello, how do you get the parameters of K, L , M. I looked at the pdf posted on the first page but I am still confused o how to get the Parameters.. can someone please help me? p.s What is regression??Hey. If you have a GDC like TI-83 or TI-84 this should be easy. Go to STATS then then click on edit and enter all the data that was given. The years are too large and makes it harder to graph so you can input 0,5,10,15,20,25,30,35,40,45. These are the exact intervals as the years so the accuracy is fine. After you enter both the year and population value quit and then go back to STATS. Go across to CALC then scroll down all the way to logistic...which is B I believe then press enter then put L1 and L2. Then waalaa you have your parameters. hey what do you mean by "put L1 and L2"? i did it your way and i got c=1946.183708 b= 0.0333213657 a=2.619019043 and i graphed this function and compared with the population trends in China but they were way different. Can you help please? Reply Link to post Share on other sites More sharing options...
dessskris Posted February 6, 2011 Report Share Posted February 6, 2011 hey what do you mean by "put L1 and L2"? i did it your way and i got c=1946.183708 b= 0.0333213657 a=2.619019043 and i graphed this function and compared with the population trends in China but they were way different. Can you help please?What GDC do you have? Read the quote below vHow to Use Technology to Generate A FunctionThis procedure below are meant for TI-Nspire™ calculators.1. Add Lists & Spreadsheet2. Put in the values of your variables (time and population) in column A and B3. Insert a new page4. Add Data & Statistics5. Plot Properties > Add X Variable > choose your independent variable (time)6. Plot Properties > Add Y Variable > choose your dependent variable (population)7. Data series will appear8. Analyze > RegressionShow Linear (mx+b)Show Linear (a+bx)Show Median - MedianShow QuadraticShow CubicShow QuarticShow PowerShow ExponentialShow LogarithmicShow SinusoidalShow Logistic (d=0)Show Logistic (d≠0)9. Pick one desired regression10. Trendline will appear and the equation will also be shown11. You are doneI hope the procedure is similar for the other TI calculators. Reply Link to post Share on other sites More sharing options...
greenglob Posted February 6, 2011 Report Share Posted February 6, 2011 Um this is mainly directed to Desy (Sorry I'm not completely sure on how to quote someone on this forum)I understand that we need to use regressions to develop the Logistic equation but almost everyone in my class knows about the Logistic model. So I was wondering what is the "wow" factor for this portfolio? I thought if I was able to set up the equation algebraically, I'd be able to get the marks but it's hella hard. Any info would be appreciated.Also I'm kind of confused about substituting the values for the the parameters and the variable... Reply Link to post Share on other sites More sharing options...
1-2-3 Posted February 6, 2011 Report Share Posted February 6, 2011 (edited) Hi can anyone help me?1) For the part where you have to analytically develop two functions that model the data points, people have been saying the cubic and expo work the best? ( I think)However, the previous question requires a justification for the functions that you are going to create and why it would be a good one to use. Expo is easy to think of, but why cubic? I don't understand why everyone is using cubic!! I was playing with the trendlines in excel and it seems that the higher power polynomial functions fit a bit better...why are people using cubic?!?2) ALSO: Does anyone know how to create an exponential function algebraically? Desy, thanks so much for you scan! It helped a lot for the polynomials, could you teach me how to do it with exponential? (I'm pretty sure you can only use polynomial functions with matrices to solve)ThanksAlso there is one thing I don't understand.Desy, you know how you substitued numbers 0, 1, 2.... for the years 1950, 1955.... when you were solving for the polynomial equations by hand with matrices?How does that work? I mean if you change the x values that will change your entire graph! Does that mean you have to graph EVERYTHING from that point on with the years as 0, 1, 2, 3....instead of 1950, 1955,1960....? Edited February 7, 2011 by Desy ♫ Reply Link to post Share on other sites More sharing options...
Hmillz Posted February 6, 2011 Report Share Posted February 6, 2011 Has anybody finished yet? and how should you do: construct the researcher's model using your estimates.What estimates???? Reply Link to post Share on other sites More sharing options...
dessskris Posted February 7, 2011 Report Share Posted February 7, 2011 sorry people I'm taking years to reply! I was doing my own HW, still haven't finished it but anyways..Um this is mainly directed to Desy (Sorry I'm not completely sure on how to quote someone on this forum)I understand that we need to use regressions to develop the Logistic equation but almost everyone in my class knows about the Logistic model. So I was wondering what is the "wow" factor for this portfolio? I thought if I was able to set up the equation algebraically, I'd be able to get the marks but it's hella hard. Any info would be appreciated.Also I'm kind of confused about substituting the values for the the parameters and the variable...You can always click the Reply button under the post that you want to quote Oic..well you may want to do an analytical approach then if you feel like it.So you have:P = K / (1 + L e-Mt) where K, L, M are parameters.You need to solve for 3 parameters. So you need 3 data points. Just pick from your data! Since I'm giving an example I won't use the original data ok.Let's say we have (5, 10), (10, 20), (15, 30) btw format is (t, P)Substituting those values into the general equation you have:10 = K / (1 + L e-5M)20 = K / (1 + L e-10M)30 = K / (1 + L e-15M)From 1st eq:L e-5M = K/10 - 1Do the same for the rest and just try to solve idk though I haven't tried that either! If you dare to try but later find some difficulties, reply here and I will try to help Hi can anyone help me?1) For the part where you have to analytically develop two functions that model the data points, people have been saying the cubic and expo work the best? ( I think)However, the previous question requires a justification for the functions that you are going to create and why it would be a good one to use. Expo is easy to think of, but why cubic? I don't understand why everyone is using cubic!! I was playing with the trendlines in excel and it seems that the higher power polynomial functions fit a bit better...why are people using cubic?!?2) ALSO: Does anyone know how to create an exponential function algebraically? Desy, thanks so much for you scan! It helped a lot for the polynomials, could you teach me how to do it with exponential? (I'm pretty sure you can only use polynomial functions with matrices to solve)ThanksAlso there is one thing I don't understand.Desy, you know how you substitued numbers 0, 1, 2.... for the years 1950, 1955.... when you were solving for the polynomial equations by hand with matrices?How does that work? I mean if you change the x values that will change your entire graph! Does that mean you have to graph EVERYTHING from that point on with the years as 0, 1, 2, 3....instead of 1950, 1955,1960....?1. Yeah 6th degree polynomial fits much better I don't know why either but maybe because it's easier to model..? Actually there is a trigonometrical function that fits really well, much better than cubic I would say. Try to find it if you want.. I believe it has been mentioned in this thread too 2. You are welcome! You are actually asked to model one function only but up to you.. I actually don't see how exponential fits, really it doesn't fit that well, I guess. But if you still want to model exponential, you start with the general formula first.P = a + bctI have not much idea but perhaps same as logistic (see my reply to greenglob), substitute values of t and P and solve for a, b, c using elimination and substitution if you face any problem let me know, will try to help you out!- Manipulate the function just make up a new variable whose values will be 1, 2, 3, ...Let's say your t are 1940, 1942, 1944, 1946, 1948 and 1950 and you want to change it to small numbers,Let x=(t-1940)/2Get that? But dont forget in your final function it must be P in terms of t and not just x so substitute it back.I hope it's clear enough :/ if still unclear just ask again!Has anybody finished yet? and how should you do: construct the researcher's model using your estimates.What estimates????I'm not doing this task just want to help people around..Use technology to estimate those parameters first! As I said, plot the graph and find a logistic regression and that's your estimates of K, L and M from technology. That makes your researcher's model don't forget to interpret those parameters btw! 1 Reply Link to post Share on other sites More sharing options...
fudger562 Posted February 8, 2011 Report Share Posted February 8, 2011 I have another question. So I modeled a cubic function and it was very close to the original graph. Our teacher tells us about some regression thing on our TI-83+ and it gives me a different answer. I compared the data from the function the calculator gave us and the original data and it was much further off than what i had. So can someone explain to me what's going on? Does this mean I did the entire portfolio wrong? Reply Link to post Share on other sites More sharing options...
Daniel Inchan Jung Posted February 8, 2011 Report Share Posted February 8, 2011 Has anybody finished yet? and how should you do: construct the researcher's model using your estimates.What estimates????Obviously you use the researcher's given sigmoid function.Find the variables like you did with your own model. Reply Link to post Share on other sites More sharing options...
dessskris Posted February 8, 2011 Report Share Posted February 8, 2011 I have another question. So I modeled a cubic function and it was very close to the original graph. Our teacher tells us about some regression thing on our TI-83+ and it gives me a different answer. I compared the data from the function the calculator gave us and the original data and it was much further off than what i had. So can someone explain to me what's going on? Does this mean I did the entire portfolio wrong?I modeled a cubic function and it was very close to the original graphGood! you are awesome!some regression thing on our TI-83+ and it gives me a different answerOf course it does! Ok, so when you modelled your own function, you might have used 4 different data points to actually get your cubic function (if you did it in a similar way with what I described in my scanned pic a few pages behind). The task gives you more than 4 data points, which means you can model more than one functions using exactly the same method but with different data points. FYI, if you pick points A, B, C, D and then you pick points E, F, G, H you will get totally different cubic functions. Or even points A, B, C, D and points A, B, C, E will give you a different function already.That is why your model function might be different with your friends' functions so it is also possible to get a function other than the one generated using technology So don't worry, you are on the right track Reply Link to post Share on other sites More sharing options...
Ridamalik Posted February 9, 2011 Report Share Posted February 9, 2011 If any one needs help with this task, I finished it last term, so just ask Got 19/20 points on it.Did you use exponential or logistic graph??? Reply Link to post Share on other sites More sharing options...
ibibib1022 Posted February 9, 2011 Report Share Posted February 9, 2011 Hi guys!! My portfolio's due the day after tomorrow and I'm stuck.. I want to find a logarithmic or exponential model by hand, but I don't know how... Any help please? Reply Link to post Share on other sites More sharing options...
dessskris Posted February 9, 2011 Report Share Posted February 9, 2011 Did you use exponential or logistic graph???FYI, you actually need two graphs: logistic and another type of function (can be exponential even though I do not really suggest it)Hi guys!! My portfolio's due the day after tomorrow and I'm stuck.. I want to find a logarithmic or exponential model by hand, but I don't know how... Any help please?You should have started doing it at least 3 days ago I think that's a bit hard. State the general equation for that type of function first and then you substitute some sets of values of t and P to that general equation and then you solve them using elimination and/or substitution If you don't understand look at post #314, my reply to greenglob, a similar rough example on how to model a function.If still unclear just ask again! 1 Reply Link to post Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.