毕业设计基于机器学习python实现的重庆轨道交通客流量时空分析预测项目源码.zip

import datetime import json import pickle import random import numpy import matplotlib import numpy as np import scipy from sklearn import metrics, ensemble from sklearn.metrics import mean_absolute_error from sklearn.neighbors import RadiusNeighborsRegressor from sklearn.neural_network import MLPRegressor from sklearn.preprocessing import OneHotEncoder, MinMaxScaler import pymysql from django.core import serializers from django.http import JsonResponse from django.shortcuts import render from sklearn.svm import SVR from sklearn import linear_model from django.views.decorators.csrf import csrf_exempt # Create your views here. from django.views.decorators.csrf import csrf_exempt from django.views.decorators.http import require_http_methods from sklearn.tree import DecisionTreeRegressor from RailTransitCrowdForecasting.models import Lineflow, Stationflow, Station, Linedayflow, Stationdayflow # 获取站点信息 @require_http_methods(["GET"]) def get_line(request, ): response = {} idlist = [] lines = Lineflow.objects.values() mset = set() for line in lines: if line["lineid"] not in mset: mdic = dict() mset.add(line["lineid"]) mdic["lineid"] = line["lineid"] idlist.append(mdic) response["linelist"] = idlist return JsonResponse(response) @require_http_methods(["GET"]) def initial(request, ): response = {} stations = Station.objects.values() line_set = set() res = [] for station in stations: if station["lineid_id"] not in line_set: line_set.add(station["lineid_id"]) line = {"value": station["lineid_id"], "label": str(station["lineid_id"]) + "号线", "children": []} children = {"value": station["stationid"], "label": station["stationname"]} line["children"].append(children) res.append(line) else: for line in res: if line["value"] == station["lineid_id"]: children = {"value": station["stationid"], "label": station["stationname"]} line["children"].append(children) response["list"] = res return JsonResponse(response) # 获取历史客流量 @require_http_methods(["GET"]) def getFlow(request): lineID = request.GET["lineID"] stationID = request.GET["stationID"] date = request.GET["date"] # lineID = '3' # stationID = '303' # date = '2018-06-05' dateFormat = datetime.datetime.strptime(date, "%Y-%m-%d") lineRes = Linedayflow.objects.filter(lineid=lineID, date=dateFormat).values() stationRes = Stationdayflow.objects.filter(stationid=stationID, date=dateFormat).values() stationName = Station.objects.filter(stationid=stationID).values()[0]["stationname"] lineFlowList = [] stationFlowList = [] for line in lineRes: lineFlowList.append(line["flow"]) for station in stationRes: stationFlowList.append(station["flow"]) response = {"lineCheck": {"lineID": lineID, "lineName": str(lineID) + "号线", "date": date, "flow": lineFlowList}, "stationCheck": {"stationID": stationID, "stationName": stationName, "date": date, "flow": stationFlowList} } return JsonResponse(response) # 预测客流量 @csrf_exempt @require_http_methods(["POST"]) def predict(request): params = json.loads(request.body)["params"] lineID = params["lineID"] stationID = params["stationID"] date = params["date"] methodList = params["methods"][1:-1].replace('"', '').split(',') response = {} response["station"] = {} response["line"] = {} # lineID = '3' # stationID = '303' # date = '2018-06-03' # methodList= "BPNN,BPNN" dayType = datetime.datetime.strptime(date, "%Y-%m-%d").weekday() + 1 stationName = Station.objects.filter(stationid=stationID).values()[0]["stationname"] for method in methodList: print(method) m = datetime.timedelta(hours=7) station_pre_flow = list() line_pre_flow = list() station_test_flow = list() line_test_flow = list() while m <= datetime.timedelta(hours=23): station_data = get_station_data(m, stationID, date) line_data = get_line_data(m, lineID, date) station_res = train(station_data, method) line_res = train(line_data, method) station_test_flow.append(station_res[0]) station_pre_flow.append(station_res[1]) line_test_flow.append(line_res[0]) line_pre_flow.append(line_res[1]) m = m + datetime.timedelta(minutes=5) station_mse = metrics.mean_squared_error(station_test_flow, station_pre_flow) station_rmse = np.sqrt(station_mse) station_mae = metrics.mean_absolute_error(station_test_flow, station_pre_flow) line_mse = metrics.mean_squared_error(line_test_flow, line_pre_flow) line_rmse = np.sqrt(line_mse) line_mae = metrics.mean_absolute_error(line_test_flow, line_pre_flow) response["station"][method] = { "preflow": station_pre_flow, "mse": int(station_mse), "rmse": int(station_rmse), "mae": int(station_mae) } response["line"][method] = { "preflow": line_pre_flow, "mse": int(line_mse), "rmse": int(line_rmse), "mae": int(line_mae) } response["station"]["stationName"] = stationName response["station"]["stationID"] = stationID response["station"]["date"] = date response["station"]["testflow"] = station_test_flow response["line"]["lineName"] = str(lineID) + "号线" response["line"]["lineID"] = lineID response["line"]["date"] = date response["line"]["testflow"] = line_test_flow return JsonResponse(response) # # # bpnn预测客流量 # @require_http_methods(["GET"]) # def bpnnpredict(request): # # print(request) # # print("12345") # lineID = request.GET["lineID"] # stationID = request.GET["stationID"] # date = request.GET["date"] # # lineID = '3' # # stationID = '303' # # date = '2018-06-03' # dayType = datetime.datetime.strptime(date, "%Y-%m-%d").weekday() + 1 # stationName = Station.objects.filter(stationid=stationID).values()[0]["stationname"] # m = datetime.timedelta(hours=7) # station_test_flow = list() # station_pre_flow = list() # line_test_flow = list() # line_pre_flow = list() # while m <= datetime.timedelta(hours=23): # station_data = get_station_data(m, stationID, date) # line_data = get_line_data(m, lineID, date) # station_res = bpnn_train(station_data) # line_res = bpnn_train(line_data) # station_test_flow.append(station_res[0]) # station_pre_flow.append(station_res[1]) # line_test_flow.append(line_res[0]) # line_pre_flow.append(line_res[1]) # m = m + datetime.timedelta(minutes=5) # # station_mse = metrics.mean_squared_error(station_test_flow, station_pre_flow) # station_rmse = np.sqrt(station_mse) # station_mae = metrics.mean_absolute_error(station_test_flow, station_pre_flow) # line_mse = metrics.mean_squared_error(line_test_flow, line_pre_flow) # line_rmse = np.sqrt(line_mse) # line_mae = metrics.mean_absolute_error(line_test_flow, line_pre_flow) # response = {"line": {"lineID": lineID, "lineName": str(lineID) + "号线", "date": date, # "testflow": line_test_flow, "preflow": line_pre_flow, # "mse": int(line_mse), "rmse": int(line_rmse), "mae": int(line_mae)}, # "station": {"stationID": stationID, "stationName": stationName, "date": date, # "testflow": station_test_flow, "preflow": station_pre_flow, # "mse": int(station_mse), "rmse": int(station_rmse), "mae": int(station_mae)} # } # # return JsonResponse(response) # # #