Machine learning Curve fitting using MATLAB

here is the code for the data set we have taken.



>> clear;
close all;
Tx = [80: -20: -340];
alphay=[6.47  6.36 6.24 6.12 6.00 5.86 5.72 5.58 5.43 5.28 5.09 4.91 4.72 4.52 4.30 4.08 3.83 3.58 3.33 3.07 2.76 2.45];
alphay=1.e-6*alphay;
plot(Tx,alphay,'bo');


title('Thermal expansion cofficient of steel');
ylabel('cofficient\alpha');
xlabel('temperature(in F)');
>> coef1=polyfit(Tx,alphay,1)



coef1 =

   1.0e-05 *

    0.0009    0.6025

>> coef1=polyfit(Tx,alphay,2)

coef1 =

   1.0e-05 *

   -0.0000    0.0006    0.6015

>> coef1=polyfit(Tx,alphay,3)

coef1 =

   1.0e-05 *

    0.0000   -0.0000    0.0006    0.5998

>> pred1=polyval(coef1,Tx);
>> pred2=polyval(coef2,Tx);
Undefined function or variable 'coef2'.

Did you mean:
>> coef1=polyfit(Tx,alphay,1)

coef1 =

   1.0e-05 *

    0.0009    0.6025

>> coef2=polyfit(Tx,alphay,2)

coef2 =

   1.0e-05 *

   -0.0000    0.0006    0.6015

>> coef3=polyfit(Tx,alphay,3)

coef3 =

   1.0e-05 *

    0.0000   -0.0000    0.0006    0.5998

>> pred1=polyval(coef1,Tx);
>> pred2=polyval(coef2,Tx);
>> pred3=polyval(coef3,Tx);
>> [pred1; pred2; pred3]

ans =

   1.0e-05 *

  Columns 1 through 7

    0.6776    0.6588    0.6400    0.6213    0.6025    0.5837    0.5649
    0.6432    0.6342    0.6243    0.6134    0.6015    0.5886    0.5748
    0.6467    0.6357    0.6243    0.6123    0.5998    0.5866    0.5727

  Columns 8 through 14

    0.5462    0.5274    0.5086    0.4898    0.4711    0.4523    0.4335
    0.5599    0.5441    0.5273    0.5095    0.4907    0.4710    0.4502
    0.5581    0.5426    0.5264    0.5092    0.4910    0.4719    0.4517

  Columns 15 through 21

    0.4147    0.3960    0.3772    0.3584    0.3397    0.3209    0.3021
    0.4285    0.4058    0.3821    0.3574    0.3318    0.3052    0.2776
    0.4303    0.4078    0.3841    0.3592    0.3328    0.3052    0.2760

  Column 22

    0.2833
    0.2490
    0.2455

>> hold on
>> plot(tx,pred1)
Undefined function or variable 'tx'.

Did you mean:
>> plot(Tx,pred1)
>> plot(Tx,pred2,'r')
>> plot(Tx,pred3,'k--')
>> legend({'Data' 'Linear' 'Quadratic' 'Cubic'}, 'Location', 'Northwest')
>> hold off
>>