### New Square Method

Key Words: multi-dimensional, non-linear, data regression, model, application

**ABSTRACT**

The “new square method” is an improved approach based on the “least square method”. It calculates not only the constants and coefficients but also the variables’ power values in a model in the course of data regression calculations, thus bringing about a simpler and more accurate calculation for non-linear data regression processes.

**Preface**In non-linear data regression calculations, the “least square method” is applied for mathematical substitutions and transformations in a model, but the regression results may not always be correct, for which we have made improvement on the method adopted and named the improved one as “new square method”.

**Principle of New Square Method**While investigating the correlation between variables(x,y), we get a series of paired data (x

_{1},y_{1},x_{2},y_{2}……x_{n},y_{n}) through actual measurements. Plot these data on the x – y coordinates, then a scatter diagram as shown in Figure 1 will be obtained. It can be observed that the points are in the vicinity of a curve, whose fitted equation is set as the following Equation 1.

Figure 1

Equation 1where a

_{0}, a_{1}and k indicate any real numbers.To establish the fitted equation, the values of a0, a1 and k need to be determined via subtracting the calculated valuefrom the measured value y

_{i}, i.e., via(y_{i}- y).Then calculate the quadratic sum of m(y

_{i}- y) as shown in Equation 2.Equation 2Substitute Expression 1 into Expression 2, as shown in Expression 3:

Equation 3Find the partial derivatives for a0, a1 and k respectively through functionso as to make the derivatives equal to zero:

Equation 4Equation 5Equation 6Through derivation it is found that there is no analytic solution to this equation set, then computer programs are utilized to calculate its arithmetic solutions and obtain the solutions for a

_{0}, a_{1}and k as well as the correlation coefficient. It is observed that the closer the correlation coefficient is to 1, the better the model fits.**Comparison between the “New Square Method” and the “Least Square Method”**If Equation 7 as shown below is adopted to fit any data:

y = a_{0}+ a_{1}x_{i}^{k}Equation 7The Comparison Table between the “New Square Method” and the “Least Square Method”:

Least Square Method New Square Method Fitted Equations y = a _{0}+ a_{1}xy = a _{0}+ a_{1}x_{i}^{k}Calculated Regression Results a _{0}and a_{2}a _{0}and a_{2}and k- In the “new square method”, the power valueof the dependent variable is calculated, while in the “least square method”, is assumed to be 1. With the calculated power value for the dependent variable, the new square method is able to have the fitted equation generate a fitted line at any curve to better fit the non-linear data.
- In the “new square method”, non-linear data with one factorcan be regressed by applying the following Equation 8 in the computer programs to obtain more accurate fittings of non-linear data by regression models.
y = a
_{0}+ a_{1}x^{k1}+ a_{2}x^{k2}+......+a_{n}x^{kn}Equation 8In Equation 8:

x — Variable

y — Variable

x,y — Dimensional (two-dimensional).

x^{k1},x^{k2},x^{kn}— Element.

a_{0}— Constant.

a_{1}, a_{2}, a_{n}— Coefficient.

k1, k2, kn — Power. - As for the regression of non-linear data with multi-factors in the “new square method”, the following Equation 9 can be utilized in computer programs for this purpose. This equation takes into account both the contribution of factors(x
_{1}, x_{2}.....x_{n}) to the objective function (y) and the interplays among factors(x_{1}, x_{2}.....x_{n}) during the regression calculation, that is why the fitted models are of high correlation.y = a_{0}+ a_{1}x_{1}^{k11}+ a_{2}x_{2}^{k21}+ a_{3}x_{1}^{k12}x_{2}^{k22}+ a_{4}x_{1}^{k13}x_{2}^{k23}+......+a_{n+1}x_{1}^{k1n+1}x_{2}^{k2n+1}Equation 9In Equation 9:

x_{1}, x_{2}— Variable.

y — Function.

x_{1},x_{2},y — Dimensional (three-dimensional).

x_{1}^{k11}, x_{2}^{k21},x_{1}^{k12},x_{2}^{k22}, x_{1}^{k13}, x_{2}^{k23}, x_{1}^{k1n+1}, x_{2}^{k2n+1}— Element.

a_{0}, — Constant.

a_{1}, a_{2}, a_{3}, a_{4}, a_{n+2}— Coefficient.

k11, k21, l12, k22, k13, k23, k1n + 1, k2n + 1 — Power.**Note:**Equation 9, which takes three-dimensional data as its example, can be applied for the regression of data in curved surface data.**Bibliography:**- Shoupeng Wei. 1994. Petrochemical production process optimization. Sinopec press.
- Chengsen Lin. 1997. The numerical calculation method. China Science Press.
- Shisong Mao and Jixiang Zhou.1996. The theory of probability and statistics. China Statistics Press.
- Jeffrey Ri hter.2003. Microsoft.NET framework of the program design. Tinghua University press.

### Comments:

10 Mar 2019 10:10:37 AM

10 Mar 2019 04:54:56 PM

06 Mar 2019 03:24:09 PM