public void test_AdjustedR2_performance_10000_observations_2_variables()
{
Stopwatch sw = new Stopwatch();
Random rand = new Random();
List <double> y = new List <double>();
List <double> x1 = new List <double>();
List <double> x2 = new List <double>();
for (int i = 0; i < 10000; i++)
{
y.Add(rand.Next(0, 10));
x1.Add(rand.Next(0, 10));
x2.Add(rand.Next(0, 10));
}
List <List <double> > expl = new List <List <double> >()
{
x1, x2
};
sw.Start();
ModelFit.AdjustedR2(y, expl);
sw.Stop();
Console.WriteLine("Took " + sw.ElapsedMilliseconds + " milliseconds");
Assert.LessOrEqual(sw.ElapsedMilliseconds, 1000);
}
public void test_AdjustedR2_return_correct_result()
{
/*Calculated with R, lm (list ~ list2 + list3) produces an adjusted R2 value of 0.2081.*/
List <List <double> > explanatory = new List <List <double> >()
{
list2, list3
};
Assert.AreEqual(0.2081, Math.Round(ModelFit.AdjustedR2(list, explanatory), 4));
}
public void test_AdjustedR2_throws_MathError_Different_Sized_Lists()
{
/*Different sized lists throw MathError*/
List <double> y = new List <double>()
{
2, 3
};
List <List <double> > lists = new List <List <double> >()
{
list, list2, list3
};
Assert.Throws <MathError>(() => ModelFit.AdjustedR2(y, lists));
}
public void test_FittedValues_throws_MathError()
{
/*If any explanatory variable list is different size than y.Count, MathError is thrown.*/
List <double> y = new List <double>()
{
2, 3
};
List <List <double> > lists = new List <List <double> >()
{
list, list2, list3
};
Assert.Throws <MathError>(() => ModelFit.FittedValues(y, lists));
}
public void test_FittedValues()
{
/*Tested on R, a model lm(list ~ list2 + list3)
* gives fitted values of 0.9818711, 3.5873076, 5.7066397, 8.1032177, 10.5016440, 3.8193199 */
List <double> expected = new List <double>()
{
0.9818711, 3.5873076, 5.7066397, 8.1032177, 10.5016440, 3.8193199
};
List <List <double> > explanatory = new List <List <double> >()
{
list2, list3
};
List <double> results = ModelFit.FittedValues(list, explanatory);
results = results.Select(y => Math.Round(y, 7)).ToList();
Assert.AreEqual(expected, results);
}
public void test_AIC()
{
/*Result calculated with R.*/
List <double> y = new List <double>()
{
1, 2.2, 3.1, 2.5
};
List <double> x1 = new List <double>()
{
177, 175, 183, 167
};
List <double> x2 = new List <double>()
{
3, 5, 6, 9
};
List <List <double> > x = new List <List <double> >()
{
x1, x2
};
Assert.AreEqual(-4.032542, Math.Round(ModelFit.AIC(y, x), 6));
}
public void test_AdjustedR2_throws_MathError()
{
/* Trying to get coefficients for invalid matrix should produce ArgumentException.*/
List <double> y = new List <double>();
List <double> x1 = new List <double>();
List <double> x2 = new List <double>();
List <double> x3 = new List <double>();
for (int i = 0; i < 100; i++)
{
y.Add(1);
x1.Add(1);
x2.Add(1);
x3.Add(1);
}
List <List <double> > expl = new List <List <double> >()
{
x1, x2, x3
};
Assert.Throws <MathError>(() => ModelFit.AdjustedR2(y, expl));
}
/// <summary>
/// returns the akaike information criteria value for model m.
/// </summary>
/// <exception cref="ArgumentException">Thrown when any variable in the model has
/// different number of obesrvations.</exception>
public double EvaluateFitness(Model m)
{
return(ModelFit.AIC(m.getYVar().values, m.getXVariableLists()));
}
