数学建模python实现基础编程练习4

数学建模python实现基础编程练习4

1.使用networkx实现最短路和最短距离

import networkx as nx
list1 = [(1,2,7),(1,3,3),(1,4,12),(2,3,3),(2,6,2),(3,4,8),(4,5,1),(5,6,3)]#邻接矩阵
g = nx.Graph()
g.add_weighted_edges_from(list1)
s = nx.to_numpy_matrix(g,nodelist=range(1,7))
p = nx.dijkstra_path(g,source=1,target=5,weight='weight')
d = nx.dijkstra_path_length(g,1,5,weight='weight')
print("最短路径：",p)
print("最短距离：",d)

2.生成最小树

tree = nx.minimum_spanning_tree(g,weight='weight',algorithm='prim')
w = nx.get_edge_attributes(tree,'weight')
print("最小生成树：",w)

3.在何处开小学使总路程最短

我把abcd换成1234…

import networkx as nx
list1 = [(1,2,2),(1,3,7),(2,3,4),(2,4,6),(2,5,8),(3,4,1),(3,5,3),(4,5,1),(4,6,6),(5,6,3)]
g = nx.Graph()
g.add_nodes_from(range(1,7))
g.add_weighted_edges_from(list1)
c = dict(nx.shortest_path_length(g,weight='weight'))
d = np.zeros((6,6))
for i in range(1,7):
    for j in range(1,7):
        d[i-1,j-1] = c[i][j]
print(d)
q = np.array([100,80,60,40,70,90])
m = d@q
mm = m.min()
ind =np.where(m==mm)[0]+1
print("最短总路程：",mm)
print("在{}开小学".format(ind))

4.pagerank算法

import networkx as nx
from scipy.sparse.linalg import eigs
list1 = [(1,2),(1,4),(1,5),(1,6),(2,4),(2,5),(2,6),(3,1),(3,2),(3,4),(4,5),(4,6),(5,3),(5,6),(6,3)]
g = nx.DiGraph()
g.add_nodes_from(range(1,7))
g.add_edges_from(list1)
b = np.array(nx.to_numpy_matrix(g))
a = b/np.tile(b.sum(axis=1,keepdims=True),(1,b.shape[1]))
a = 0.15/b.shape[0]+0.85*a
print(a)
w,v = eigs(a.T,1);v = v.real
v = v.flatten()
v = v/v.sum()
print(v)
plt.figure(2)
plt.bar(range(1,b.shape[0]+1),v,width=0.6,color='b')
plt.show()

注意的是邻接矩阵不要写反（我这里就写反了）

5.采用距离判别法进行判别分析

rom sklearn.neighbors import KNeighborsClassifier
g1 = np.array([[3,2],[2,4],[4,7]])
s1 = g1.sum(axis=0,keepdims=True)/len(g1)
c1 = np.cov(g1)
print("g1均值向量:",s1[0])
print("g1协方差:",c1)
g2 = np.array([[6,9],[5,7],[4,8]])
s2 = g2.sum(axis=0,keepdims=True)/len(g2)
c2 = np.cov(g2)
print("g2均值向量:",s2[0])
print("g2协方差:",c2)

x0 = np.array([[3,2],[2,4],[4,7],[6,9],[5,7],[4,8]])
y = np.array([1,1,1,2,2,2])
v = np.cov(x0.T)
k = KNeighborsClassifier(2,metric='mahalanobis',metric_params={'V':v})
k.fit(x0,y)
xx = np.array([[2,7]])
pre = k.predict(xx)
print(pre)

6.主成分分析

from sklearn.decomposition import PCA
from scipy.states import zscore
x0 = np.array([[1,0.80,0.37,0.78,0.26,0.38],[0.80,1,0.32,0.65,0.18,0.33],[0.37,0.32,1,0.36,0.71,0.62],[0.78,0.65,0.36,1,0.18,0.39],[0.26,0.18,0.71,0.18,1,0.69],[0.38,0.33,0.62,0.39,0.69,1]])
x0 =zscore(x0)#数据标准化
md = PCA()
md.fit(x0)
print("特征值：\n",md.explained_variance_)
print("各主成分的贡献值：\n",md.explained_variance_ratio_)
print("奇异值：\n",md.singular_values_)
print("各主成分的系数：\n",md.components_)

7.因子分析模型

from sklearn.decomposition import FactorAnalysis
from scipy.stats import zscore
import pandas as pd
x1 = np.array([[1,0.80,0.37,0.78,0.26,0.38],[0.80,1,0.32,0.65,0.18,0.33],[0.37,0.32,1,0.36,0.71,0.62],[0.78,0.65,0.36,1,0.18,0.39],[0.26,0.18,0.71,0.18,1,0.69],[0.38,0.33,0.62,0.39,0.69,1]])
x0 = zscore(x1)#数据标准化
r = np.corrcoef(x0.T)#求相关系数矩阵
val, vec = np.linalg.eig(r)#特征值
cs = np.cumsum(val)#求特征值的累加和
fa = FactorAnalysis(n_components=2)
fa.fit(x0)#求解模型
print("载荷矩阵：\n",fa.components_)
print("特殊方差：\n",fa.noise_variance_)
xx = fa.fit_transform(x0)#计算因子得分
w = val[:2]/sum(val[:2])#计算两个因子的权重
df = np.dot(xx,w)#计算每个评价对象的因子得分
tf = np.sum(x1,axis=1)#计算每个评价对象的实际分总分
pdf = pd.DataFrame(np.c_[xx,df,tf,np.arange(1,7)],columns=['f1','f2','yf','tf','xh'])
spdf1 = pdf.sort_values(by='yf',ascending=False)
spdf2 = pdf.sort_values(by='tf',ascending=False)
print("排序结果：\n",spdf1,'\n',spdf2)
s = ['s'+str(i) for i in range(1,7)]
plt.plot(xx[:,0],xx[:,1],'.')
for i in range(len(s)):plt.text(xx[i,0],xx[i,1]+0.03,s[i])
plt.show()