Get class output as np.array

Question

I was asking myself if it is possible to turn the output of a class into a np.array within the class itself. I created the following class:

class stats:

    def __init__( self, x ):
        self.age = x[:,0]
        self.education = x[:,1]
        self.married = x[:,2]
        self.nodegree = x[:,3]
        self.RE75 = x[:,4]
        self.RE78 = x[:,5]

    def Vector( self ):
        age = [np.mean(self.age), st.stdev(self.age)]
        education = [np.mean(self.education), st.stdev(self.education)]
        married = [np.mean(self.married), st.stdev(self.married)]
        nodegree = [np.mean(self.nodegree), st.stdev(self.nodegree)]
        RE75 = [np.mean(self.RE75), st.stdev(self.RE75)]
        RE78 = [np.mean(self.RE78), st.stdev(self.RE78)]
        return [age, education, married, nodegree, RE75, RE78]

results1 is a numpy.ndarray of shape 156x6. I basically want to compute the mean as well as standard deviation for each column of results1 using a class. I use numpy to compute the mean and statistics for the std. When I am printing the output I get the following:

results1_stats = stats(results1)
print(results1_stats.Vector())

Output:

[[25.98076923076923, 7.299554695959556], [10.314102564102564, 2.0597666237347005], [0.1858974358974359, 0.39027677820527085], [0.7243589743589743, 0.448275807219502], [1490.7220884615383, 3296.5535502409775], [6136.320646794872, 8143.4659725229685]]

Apparently, the class is working as wanted (although there is probablly a more efficent way to code this up). The problem is, that I would lilke to get the output in a np.array of shape 6x2 (or transposed) directly from the class itself. However, since I just began using classes I don't know if that is even possible. Any help is appreciated :) Thank you!

What exactly do you mean about wanting to get the output "directly from the class itself"? — martineau, Dec 11 '19 at 11:00
@martineau - Instead of getting a list for mean and std of every column, I would like to get an array where each row (or column if transposed) includes the mean & std of every column from results1 — Mcgroger, Dec 11 '19 at 11:03
@Mcgroger, in python lists are considered as arrays. Your vector function does return that list in the format you are expecting (6x2). Do you mean np.array instead of a list? perhaps, you can add expected output section in the question? — Rithin Chalumuri, Dec 11 '19 at 11:13
@Rithin Chalumuri, Yes I meant np.array. Sorry for the confusion! I have updated the question. — Mcgroger, Dec 11 '19 at 11:28

score 0 · Accepted Answer · answered Dec 11 '19 at 11:30

You can construct an numpy array using np.array(your_list_sequence). Additionally, you can use list comprehension to convert list of lists to numpy array. More info here.

Try this:

def get_stats(results):
    return np.array([np.array([np.mean(results[:, column]), st.stdev(results[:, column])]) for column in range(6)])

your_new_np_array = get_status(results)

Although, if you want only stats array, having a function for this instead of class would be better and simpler. But, you can easily include that method in your class and get back your expected result.

Get class output as np.array

1 Answers1