Beyond linear regression with pymc

Questions v5
9

Absolutely!

data_generated = pd.read_csv("data_generated.csv")
x_0 = data_generated["x0"].values
x_1 = data_generated["x1"].values
x_2 = data_generated["x2"].values
dp = np.linspace(0.1,100,len(x_0))
F = 0.6
c = []
for i in dp:
    c.append((x_1 - x_2 - x_0/F - dp) * (-F))

df_data = pd.DataFrame({"x_0" : x_0, "x1" : x_1, "x2": x_2,"C": c[1], "F" : np.zeros(len(x_0))+0.6, "dp": np.zeros(len(x_0)) + dp [1]})
df_data.to_csv("data_generated.csv", index=False)

I took c[1] which corresponds to dp = 5 and F = 0.6, x_0, x_1, x_2` are the same as their corresponding columns in the attached csv file.

My updated priors since my original post are:

    ##### priors on unknown
    C = pm.HalfNormal('coh', 25)
    F = pm.TruncatedNormal('F', mu = 0.6, sigma = 0.4, lower = 0.2)
    dp = pm.TruncatedNormal("dp", mu = 5,sigma = 6, lower = 0.1)