Internal Assessment

Mathematics: Analysis and Approaches Higher Level

Title: -To what extent can mathematical models accurately

predict India's GDP growth based on historical data

Session: - 2024 May

Introduction
Finding out how well mathematical models can forecast India's GDP growth
using historical data is the major goal of this investigation.How mathematical
modeling can reveal latent patterns and trends in economic data is a topic that
interests me.

But the issue still stands: Can mathematical models really forecast a country's
economic destiny with any degree of accuracy? To address this exact subject,
historical data analysis and a variety of modeling approaches are investigated.
In the end, this investigation extends beyond merely evaluating mathematical
models' capacity for prediction.

.Aim

The purpose of this study is to evaluate how well mathematical models

forecast GDP growth in India using historical data. Finding patterns and trends
will be accomplished by looking at different modeling techniques. The aim is to
enhance our comprehension of the predictive capacity of mathematical
models for economic fluctuations in India's complex and dynamic economy.
The ultimate goal of this research is to identify real-world uses for
mathematical modeling in economic forecasting, providing insightful
information to help policymakers and decision-makers.
Gathering Data

t-Years Since 1987 r-GDP t-Years Since 1987 r-GDP

0 283.75 19 949.12

1 299.65 20 1,238.70

2 301.23 21 1,224.10

3 326.61 22 1,365.37

4 274.84 23 1,708.46

5 293.26 24 1,823.05

6 284.19 25 1,827.64

7 333.01 26 1,856.72

8 366.60 27 2,039.13

9 399.79 28 2,103.59

10 423.19 29 2,294.80

11 428.77 30 2,651.47

12 466.87 31 2,702.93
 13 476.61 32 2,835.61

14 493.95 33 2,671.60

15 523.97 34 3,150.31

16 618.36 35 3,389.69

17 721.59 36 3,732.22

18 834.22

In table, t -0 represents the year 1987. T-1 represents the year 1988 this
continues unit t = 36 which represents 2023. From t=35 to t = 36, we can see
that this is the highest profit ever by India.
Analysis
From the data above when plotted onto a graph with the years on the x-axis
and the GDP of India made on the y-axis, we can see that there is a gradual
increase in the GDP made every year which allows the assumption that the
possible functions to model this growth would be either quadratic function or
an exponential function.

Cubic function

F(x) = 𝑎𝑥 3 + 𝑏𝑥 2 + 𝑐𝑥 + 𝑑

With this equation we are going to substitute some key features on the graph
to be able to generate an appropriate equation for the model.

Point 1 :- minimum point (0,283.75)

Point 2 :- any other point (11 , 428.77)

Point 3 :- any other point (22 ,1365.37)

Point 4 :- any other point ( 18 , 834.22)

F(x) = 𝑎𝑥 3 + 𝑏𝑥 2 + 𝑐𝑥 + 𝑑

283.75 = 𝑎(0)3 + 𝑏(0)2 + C(0) + 𝑑

283.75 = d

Point ‘ 2’ (Equation 2)

Point 2 :- Any other point (11 , 428.77)

F (x) = 𝑎𝑥 3 + 𝑏𝑥 2 + 𝑐𝑥 + 𝑑

F(x) = 𝑎(11)3 + 𝑏(11)2 + 𝑐(11) + 283.75

428.77 = 1331𝑎 + 121𝑏 + 11𝑐 + 283.75

145.02 = 1331𝑎 + 121𝑏 + 11𝑐

Point 3 :- any other point (22 ,1365.37)

F (x) = 𝑎𝑥 3 + 𝑏𝑥 2 + 𝑐𝑥 + 𝑑

F(x) = 𝑎(22)3 + 𝑏(22)2 + 𝑐(22) + 283.75

1365.37 = 10648𝑎 + 484𝑏 ⊢ 22𝑐 + 283.75

1081.62 = 10684𝑎 + 484𝑏 + 22𝑐

Point 4 :- Any other point ( 18 , 834.22)

F (x) = 𝑎𝑥 3 + 𝑏𝑥 2 + 𝑐𝑥 + 𝑑

F(x) = 𝑎(18)3 + 𝑏(18)2 + 𝑐(18) + 283.75

F(x) = 5832𝑎 + 324𝑏 + 18𝑐 + 283.75

550.47 = 5832𝑎 + 324𝑏 + 18𝑐

Solution

a = 0.1893

b= -3.0056

c= 23.3347

d= 283.75

F(x)= 0.1893𝑥 3 − 3.0056𝑥 2 + 23.3347𝑥 + 283.75

Analysis

Given that it crosses through some of the points on the graph (14 out of 36),
the graph above provides a reasonable picture of the GDP revenue statistics for
India. Even if I said that this is a decent model, it is not without flaws.Negative
years (growth decreasing) had to be disregarded in order for the cubic model to
function. This may appear to be fine, but it conceals a significant issue: the
model is unable to anticipate economic downturns or crises, which are crucial
for precise forecasting. Therefore, even though the model has its uses, it is
advisable not to rely on it in order to predict future recessions or other
significant setbacks.

From the graph it can be seen that there are some errors with the values the
cubic function calculates, and this can be seen between t = 4 and t = 12 these
values have a high percentage of error which mean that my model is
unreliable. To find the percentage error, I will be utilizing the formula:- 𝜀 =
𝑣𝑎 −𝑣𝑒
| | × 100
𝑣𝑒

𝑣𝑎 = Approximated value
𝑣𝑒 =Exact value
For the t = 12, the 𝑣𝑎 = 131 and 𝑣𝑒 = 466.87. Therefore, substituting the values
into the equation :-
131 − 466.87
𝜀=| | × 100
466.87
= 71.5%
Exponential function
General formula

𝑓(𝑡) = 𝑎𝑏𝑡 + 𝑐
derive of f(t) = ((ab)^t)+c
1. Starting with
𝐹 = 𝑎𝑡
2. Introduce a constant factor b;-
𝑔(𝑡) = (𝑎𝑏)𝑡 + 𝑐
3. Introduce an offset constant C :-
ℎ(𝑡) = (𝑎𝑏)𝑡 + 𝑐
4. Apply Derivation with Respect to ‘t’:-
𝑑ℎ 𝑑
= ((𝑎𝑏)𝑡 + 𝑐
𝑑𝑡 𝑑𝑡
Using the chain rule, the derivative of (ab)^t with respect to t is (In (ab))(ab)^t:-
𝑑ℎ
= (𝑙𝑛(𝑎𝑑))(𝑎𝑏)𝑡
𝑑𝑡
5.Integrate to find h(t) :-
To find h(t) integrate dh/dt with respect to t :-
ℎ(𝑡) = ∫ (𝑙𝑛(𝑎𝑏))(𝑎𝑏)𝑡 𝑑𝑡
The integral of (ab)^t is (ab)^t/In(ab):-
(𝑎𝑏)𝑡
ℎ(𝑡) = +𝑐
𝑙𝑛(𝑎𝑏)
Where C is the constant of integration
6. Adjust constant of integration
𝐶−1
𝐶=
𝑙𝑛( 𝑎𝑏)
−(𝑎𝑏)𝑡 +𝑐 − 1
ℎ(𝑡)
𝑙𝑛(𝑎𝑏) 𝑙𝑛(𝑎𝑏)
7. Simply and express in the form
1 𝑐−1
ℎ(𝑡) = (𝑎𝑏)𝑡 =
𝑙𝑛(𝑎𝑏) ln(𝑎𝑏)

Now let k =1/In (ab):-

ℎ(𝑡) = 𝐾(𝑎𝑏)𝑡 + 𝑐 − 𝑘
Substitute h(t) back as f(t):-
𝑓(𝑡) = 𝐾(𝑎𝑏)𝑡 + 𝑐 − 𝑘
This can be expresses in the form as :-
𝑓(𝑡) = ((𝑎𝑏)𝑡 ) + 𝑐
Since there are three variables, I will need to find three points on the graph
and substitute those values into the equation.

Definition of variables
a = The initial value
b= The growth factor or factor of decay
c= Horizontal asymptote
Point A (0,283.75)
Point B = (20,1238.70)
Point C = (36 , 3732.22)
Point A ( Equation 1 )
283.75 = ab^0+c
283.75 = a(1)+c
283.75 = a+c
c= 283.75-a
Point B ( Equation 2 )
1238.70 = 𝑎𝑏20 + 𝑐
We can replace C with Equation 1
1238.70 = 𝑎𝑏20 + (283.75 − 𝑎)
1238.70 = 𝑎𝑏20 + 283.75 − 𝑎
954.95 = 𝑎(𝑏20 − 1)
954.95
𝑎=
(𝑏20 − 1)

Point C (Equation C)
3732.22 = 𝑎𝑏25 + 𝑐
Replace ‘c’ with the equation derived in Equation 1
3732.22 = 𝑎𝑏25 + (283.75 − 𝑎)
3732.22 = 𝑎𝑏36 + 283.75 − 𝑎
3448.47 = 𝑎(𝑏36 − 1)
3448.47
a = (𝑏36
−1)
3448.47 954.95
=
(𝑏 −1) (𝑏 20 −1)
36

3448.47(𝑏20 − 1) = 954.95 ( 𝑏36 − 1)

Now both equations 2 and 3 are equal to a so we can equate them to find the
value of b
3448.47(𝑏20 − 1) = 954.95 ( 𝑏36 − 1)
By using GDC b = 1.0694
3448.47
a = (𝑏36
−1)

a = 338.23
c= -54.48
Therefore, the final equation for the exponential graph is shown below:
f(x) = 𝟑𝟑𝟖. 𝟐𝟑 × 𝟏. 𝟎𝟔𝟗𝟒𝒙 − 𝟓𝟒. 𝟒𝟖

Analysis

The exponential function I have identified has the drawback of not appearing
to have a limit or drops with the GDP made, which makes it unrealistic and
unable to be used as a model. In real life, there may be drops in the GDP made
in certain years, which may be caused by external factors like inflation. This is
evident from the graph above, which is also fairly similar to the cubic function
that was derived earlier. The above analysis leads us to the conclusion that the
exponential model is not a good fit for the data on India's GDP growth. This is
because the exponential model grows at a rate that increases infinitely, which
is not what we would find in the real world.

Quadratic graph

Point A (0,283.75)
Point B (20,1238.70)

Point C (36 , 3732.22)

F(x) = 𝒂𝒙𝟐 + 𝒃𝒙 + 𝒄

Point 1

F(0) = 𝒂(𝟎)𝟐 + 𝒃(𝟎) + 𝟐𝟖𝟑. 𝟕𝟓

C = 283.75

Point B (20,1238.70)
F(x) = 𝒂𝒙𝟐 + 𝒃𝒙 + 𝒄

1238.70 = 𝒂(𝟐𝟎)𝟐 + 𝒃(𝟐𝟎) + 𝟐𝟖𝟑. 𝟕𝟓

954.95 = 400a + 20b

954.95−400𝑎
=𝑏
20

Point C (36 , 3732.22)

3732.22= 𝑎𝑥 2 + 𝑏𝑥 + 𝑐

3732.22 = = 𝑎(36)2 + 𝑏(36) + 283.75

3448.47 = 1296a + 36b

954.95−400𝑎
3448.47 = 1296a + 36 ( )
20

a = 3.0027

b = -12.3065

c = 𝟐𝟖𝟑. 𝟕𝟓
Final Quadratic Equation: y = 3.0027𝑥 2 − 12.3065𝑥 + 283.75

Analysis
In the quadratic model, we can see that It passed through most of the data
points this also show that the quadratic function closely align with the actual
GDP data and show a more reliable model .

Standard Error

In order to determine which model, out of the ones I had previously created,
was the most accurate in terms of the difference between the actual and
predicted amounts of income, I had to provide evidence for my choice.
∑(𝑦2 −𝑦1 )2
The formula I used to calculate the standard error :- √
𝑛

𝑦2 = 𝑇ℎ𝑒 𝑒𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑟𝑒𝑣𝑒𝑛𝑢𝑒 (𝑢𝑠𝑖𝑛𝑔 𝑡ℎ𝑒 𝑒𝑞𝑢𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑚𝑜𝑑𝑒𝑙)

𝑦1 = 𝑇ℎ𝑒 𝑎𝑐𝑡𝑢𝑎𝑙 𝑟𝑒𝑣𝑒𝑛𝑢𝑒 ( 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑑𝑎𝑡𝑎 )

n = The total amount of years from 1987

Table showing the standard error of the data points of the cubic model

x 𝑦1 𝑦2 (𝑦2 − 𝑦1 ) (𝑦2 − 𝑦1 )2 (𝑦2 − 𝑦1 )2

𝑛
0 283.75 283.75 0 0 0
1 299.65 304.2684 4.6184 21.3296 0.5764
2 301.23 319.9114 18.6814 348.9947 9.4322
3 326.61 331.8148 5.2048 27.0899 0.7321
4 274.84 341.1144 66.2744 4392.2960 118.7107
5 293.26 348.946 55.686 3100.9305 83.8089
6 284.19 356.4454 72.2554 5220.8428 141.1038
7 333.01 364.7484 31.7384 1007.3260 27.2250
8 366.6 374.9908 8.3908 70.4055 1.9028
9 399.79 388.3084 -11.4816 131.8271 3.5628
10 423.19 405.8370 -17.353 301.1266 8.1385
11 428.19 428.7124 0.5224 0.2729 0.007375
12 466.87 458.0704 -8.7996 77.4329 2.0927
13 476.61 495.0468 18.4368 339.9155 9.1869
14 493.95 540.7774 46.8274 2192.8053 59.2650
15 523.97 596.3981 72.4281 5245.8296 141.7791
16 618.36 663.0444 44.6844 1996.6956 53.9647
17 721.59 741.8524 20.2624 410.5648 11.0936
18 834.22 833.9578 -0.2622 0.0687 0.0018
19 949.12 940.4964 -8.6236 74.3664 2.0099
20 1238.70 1062.6040 -176.096 31009.8012 838.1027
21 1224.10 1201.4164 -22.6836 514.5457 13.9066
22 1365.37 1358.0694 -7.3006 53.2987 1.4405
23 1708.46 1533.6888 -174.7712 30544.9723 825.539
24 1823.05 1729.4404 -93.6096 8762.7572 236.8312
25 1827.64 1946.43 118.79 14111.0641 381.3801
26 1856.72 2185.8034 329.0834 108295.8842 2926.9157
27 2039.13 2448.6964 409.5664 167744.636 4533.6388
28 2103.59 2736.2448 632.6548 400252.096 10817.6242
29 2294.80 3049.5843 754.7843 569699.3395 15397.2794
30 2651.47 3389.851 738.381 613685.7912 16586.1024
31 2702.93 3758.1804 1055.2504 1113553.407 30095.4974
32 2835.61 4155.7084 1320.0984 1742659.786 47098.9131
33 2671.60 4583.5708 1911.9708 3655632.34 98800.8740
34 3150.31 5042.9034 1892.5934 3581909.778 96808.3723
35 3389.69 5534.8421 2145.1521 4601677.532 124369.663
36 3732.22 6060.5224 2328.3024 5420992.066 146513.2991
∑(𝑦2 − 𝑦1 )2 596919.9738
𝑛
772.659
∑(𝑦2 − 𝑦1 )2
√
𝑛

Table showing the standard error of the data points of the Exponential model

x 𝑦1 𝑦2 (𝑦2 − 𝑦1 ) (𝑦2 − 𝑦1 )2 (𝑦2 − 𝑦1 )2

𝑛
0 283.75 283.75 0 0 0
1 299.65 307.2231 7.5731 57.3518 1.55
2 301.23 332.3253 31.0953 966.9176 26.1329
3 326.61 359.1 32.49 1055.6001 28.5297
4 274.84 387.8769 113.0369 12777.3407 345.3335
5 293.26 418.5765 125.3165 15704.2251 424.4385
6 284.19 451.4066 167.2166 27961.3913 755.7132
7 333.01 486.5151 153.5051 23563.8157 636.8598
8 366.6 524.0602 187.4602 35141.3265 949.7655
9 399.79 564.2109 164.4209 27034.2323 730.6549
10 423.19 607.1480 183.9580 33840.5457 914.6093
11 428.19 653.0650 134.875 18191.2656 491.6558
12 466.87 702.1686 235.2986 55365.4311 1496.3630
13 476.61 754.6801 278.0701 77322.9805 2089.8102
14 493.95 810.8357 316.8857 100416.5469 2713.9607
15 523.97 870.8887 346.9187 120352.5844 3252.7725
16 618.36 935.1093 316.7493 100330.1191 2711.6248
17 721.59 1003.7868 282.1968 79635.0339 2152.2982
18 834.22 1077.2305 243.0105 59054.1031 1596.0568
19 949.12 1155.7712 206.6512 42704.7184 6100.6740
20 1238.70 1239.7627 1.0627 1.1293 0.0305
21 1224.10 1329.5831 105.4831 11126.6843 300.7211
22 1365.37 1425.6371 60.2671 3632.1233 98.1654
23 1708.46 1528.3572 -180.1028 32437.0185 876.6761
24 1823.05 1638.2061 -184.8439 34167.2673 923.4396
25 1827.64 1755.6785 -71.9615 5178.4574 139.9583
26 1856.72 1881.3035 15.5835 242.8454 6.5633
27 2039.13 2015.6469 -23.4831 551.4559 14.9042
28 2103.59 2159.3137 55.7237 3105.1307 83.9224
29 2294.80 2312.9510 18.151 329.4588 8.9042
30 2651.47 2477.2507 -174.2193 30352.3644 820.3341
31 2702.93 2652.9528 -49.9772 2497.7205 67.5059
32 2835.61 2840.8487 5.2387 27.9174 0.7545
33 2671.60 3041.7845 370.1845 137036.564 3703.6909
34 3150.31 3256.6653 106.3553 113311.4498 305.7148
35 3389.69 3486.4587 96.7887 9364.1813 253.0859
36 3732.22 3732.1999 -0.0201 0.0004 0.00001
∑(𝑦2 − 𝑦1 )2 35023.2286
𝑛
187.1445
∑(𝑦2 − 𝑦1 )2
√
𝑛

Table showing the standard error of the data points of the Quadratic model

x 𝑦1 𝑦2 (𝑦2 − 𝑦1 ) (𝑦2 − 𝑦1 )2 (𝑦2 − 𝑦1 )2

𝑛
0 283.75 283.75 0 0 0
1 299.65 274.4462 -25.2038 635.2315 17.1684
2 301.23 271.1470 -30.083 904.9868 24.4591
3 326.61 273.8548 -52.7552 27383.1111 75.2192
4 274.84 282.5672 7.7272 59.7096 1.6137
5 293.26 297.285 4.025 16.2006 0.4378
6 284.19 318.0081 33.8181 1143.6638 30.9098
7 333.01 344.7368 11.7268 137.5178 3.7166
8 366.6 377.4708 10.8708 118.1742 3.1939
9 399.79 416.2102 16.4202 269.6229 7.2871
10 423.19 460.9550 37.765 1426.1952 38.5458
11 428.19 511.7052 83.5152 6974.7886 188.5078
12 466.87 568.4608 101.5908 10320.6906 278.9375
13 476.61 631.2218 154.6118 23904.8087 646.0759
14 493.95 699.9882 206.0382 42451.7398 1147.3443
15 523.97 774.760 250.79 62895.6241 1699.8817
16 618.36 855.5372 237.1772 56253.0242 1520.352
17 721.59 942.3198 220.7298 48721.6446 1316.8012
18 834.22 1035.1078 200.8878 40355.9081 1090.7002
19 949.12 113.9012 184.7812 34144.0918 922.8132
20 1238.70 1238.70 0 0 0
21 1224.10 1349.5042 125.4042 15726.2133 425.0327
22 1365.37 1466.3138 100.9438 10189.65076 275.3959
23 1708.46 1589.1288 -119.3312 14239.9352 384.8631
24 1823.05 1717.9492 -105.1008 11046.1781 298.5453
25 1827.64 1852.775 25.135 631.7682 17.0748
26 1856.72 1993.6062 136.8862 18737.8317 506.4278
27 2039.13 2140.4428 101.3128 10264.2834 277.4130
28 2103.59 2293.2848 189.6948 35984.1171 972.5437
29 2294.80 2452.1322 157.3322 25753.4211 669.0113
30 2651.47 2616.9850 -34.485 1189.2152 32.1409
31 2702.93 2787.8432 84.9132 7210.2515 194.8716
32 2835.61 2964.7068 129.0968 16665.9837 450.4319
33 2671.60 3147.5758 475.9758 226552.9622 6123.0530
34 3150.31 3336.4502 186.1402 34648.1740 936.4371
35 3389.69 3531.3300 141.64 20061.8896 542.2132
36 3732.22 3732.2152 0.0048 0.000023 0.000000621
∑(𝑦2 − 𝑦1 )2 21119.42
𝑛
145.3252
∑(𝑦2 − 𝑦1 )2
√
𝑛

The standard errors for all models have been calculated in conclusion Quadratic
model has the lowest standard compared to other models the standard error
for Quadratic model is 145.3252.

Evaluation

I explored multiple models to represent India GDP Growth. First I did a cubic
function in which it only fit a limited portion of the data. While the exponential
function gave a good fit in the starting but it prediction of continuous growth
didn’t reflect in the real world. Finally I did a quadratic function in which a
larger portion of data points fit and it had the lowest standard error (145.3235)
compared to other models. Therefore based on my analysis and understanding
of the actual data the quadratic function appears to be the best model for
representing India GDP growth among the three models I have explored

Conclusion

To sum up, this investigation used past data to anticipate India's GDP growth
using mathematical models. Through the analysis of cubic , exponential and
quadratic functions and seeing which models is the most accurate
representation of economic trends.
The cubic function in which had a critical limitation as it passed through (14 out
of 36 ) Data points and this function has failed to show the negative growth
years which decrease it’s ability to predict economic downturns.
The exploration of the exponential function showed that it lack the ability to
show the limit and to show the drops in the GDP. When considering external
factors like inflation. The infinite growth rate presented by the exponential
model can’t relate with the real world economic .
The quadratic function is the most suitable model for India GDP growth after
compared to the other models. It passed through most of the data points and
this model has the lowest standard error (145.3235) among the other explored
models. This show that the quadratic function closely align with the actual GDP
data and show a more reliable model
In the end the choice of the quadratic model it show a more accurate models
compared to other models. While mathematical models have limitations the
quadratic function has a lower standard error this show a good understanding
of India GDP growth patterns.

Bibliography
Calculator Suite - GeoGebra. www.geogebra.org/calculator.

“Gross Domestic Product (GDP) in India 2028.” Statista, 9 Nov. 2023,

www.statista.com/statistics/263771/gross-domestic-product-gdp-in-
india/#:~:text=Gross%20domestic%20product%20(GDP)%20in%20India%202028&t
ext=The%20statistic%20shows%20GDP%20in,the%20Russian%20GDP%20for%20c
omparison.