LR&RF.ipynb

{"cells":[{"cell_type":"markdown","source":["## Data analysis on ***\"Big-Five Factor Markers\"*** \n### 1.Load the processed dataset"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"1b66aea9-cf14-4842-ad8c-e1bb85f80ea9"}}},{"cell_type":"code","source":["file_location = \"/FileStore/tables/data_processed.csv\""],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"05eb1761-8669-4f33-b13e-087896a6bda2"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["from pyspark.sql import SparkSession\nspark = SparkSession.builder.appName('ml-bank').getOrCreate()\ndf = spark.read.csv(file_location, header = True)\ndf.printSchema()"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"2ff9395f-45fe-4533-92b8-9b52325105ad"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">root\n |-- EXT1: string (nullable = true)\n |-- EXT2: string (nullable = true)\n |-- EXT3: string (nullable = true)\n |-- EXT4: string (nullable = true)\n |-- EXT5: string (nullable = true)\n |-- EXT6: string (nullable = true)\n |-- EXT7: string (nullable = true)\n |-- EXT8: string (nullable = true)\n |-- EXT9: string (nullable = true)\n |-- EXT10: string (nullable = true)\n |-- EST1: string (nullable = true)\n |-- EST2: string (nullable = true)\n |-- EST3: string (nullable = true)\n |-- EST4: string (nullable = true)\n |-- EST5: string (nullable = true)\n |-- EST6: string (nullable = true)\n |-- EST7: string (nullable = true)\n |-- EST8: string (nullable = true)\n |-- EST9: string (nullable = true)\n |-- EST10: string (nullable = true)\n |-- AGR1: string (nullable = true)\n |-- AGR2: string (nullable = true)\n |-- AGR3: string (nullable = true)\n |-- AGR4: string (nullable = true)\n |-- AGR5: string (nullable = true)\n |-- AGR6: string (nullable = true)\n |-- AGR7: string (nullable = true)\n |-- AGR8: string (nullable = true)\n |-- AGR9: string (nullable = true)\n |-- AGR10: string (nullable = true)\n |-- CSN1: string (nullable = true)\n |-- CSN2: string (nullable = true)\n |-- CSN3: string (nullable = true)\n |-- CSN4: string (nullable = true)\n |-- CSN5: string (nullable = true)\n |-- CSN6: string (nullable = true)\n |-- CSN7: string (nullable = true)\n |-- CSN8: string (nullable = true)\n |-- CSN9: string (nullable = true)\n |-- CSN10: string (nullable = true)\n |-- OPN1: string (nullable = true)\n |-- OPN2: string (nullable = true)\n |-- OPN3: string (nullable = true)\n |-- OPN4: string (nullable = true)\n |-- OPN5: string (nullable = true)\n |-- OPN6: string (nullable = true)\n |-- OPN7: string (nullable = true)\n |-- OPN8: string (nullable = true)\n |-- OPN9: string (nullable = true)\n |-- OPN10: string (nullable = true)\n |-- country: string (nullable = true)\n\n</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">root\n-- EXT1: string (nullable = true)\n-- EXT2: string (nullable = true)\n-- EXT3: string (nullable = true)\n-- EXT4: string (nullable = true)\n-- EXT5: string (nullable = true)\n-- EXT6: string (nullable = true)\n-- EXT7: string (nullable = true)\n-- EXT8: string (nullable = true)\n-- EXT9: string (nullable = true)\n-- EXT10: string (nullable = true)\n-- EST1: string (nullable = true)\n-- EST2: string (nullable = true)\n-- EST3: string (nullable = true)\n-- EST4: string (nullable = true)\n-- EST5: string (nullable = true)\n-- EST6: string (nullable = true)\n-- EST7: string (nullable = true)\n-- EST8: string (nullable = true)\n-- EST9: string (nullable = true)\n-- EST10: string (nullable = true)\n-- AGR1: string (nullable = true)\n-- AGR2: string (nullable = true)\n-- AGR3: string (nullable = true)\n-- AGR4: string (nullable = true)\n-- AGR5: string (nullable = true)\n-- AGR6: string (nullable = true)\n-- AGR7: string (nullable = true)\n-- AGR8: string (nullable = true)\n-- AGR9: string (nullable = true)\n-- AGR10: string (nullable = true)\n-- CSN1: string (nullable = true)\n-- CSN2: string (nullable = true)\n-- CSN3: string (nullable = true)\n-- CSN4: string (nullable = true)\n-- CSN5: string (nullable = true)\n-- CSN6: string (nullable = true)\n-- CSN7: string (nullable = true)\n-- CSN8: string (nullable = true)\n-- CSN9: string (nullable = true)\n-- CSN10: string (nullable = true)\n-- OPN1: string (nullable = true)\n-- OPN2: string (nullable = true)\n-- OPN3: string (nullable = true)\n-- OPN4: string (nullable = true)\n-- OPN5: string (nullable = true)\n-- OPN6: string (nullable = true)\n-- OPN7: string (nullable = true)\n-- OPN8: string (nullable = true)\n-- OPN9: string (nullable = true)\n-- OPN10: string (nullable = true)\n-- country: string (nullable = true)\n\n</div>"]}}],"execution_count":0},{"cell_type":"code","source":["print((df.count(), len(df.columns)))"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"7b3cb6b7-ff03-4ff9-9a7c-059f90740291"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">(1012050, 51)\n</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">(1012050, 51)\n</div>"]}}],"execution_count":0},{"cell_type":"markdown","source":["### 2.Processing data"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"4069773d-4012-4504-a199-3b16faf5796f"}}},{"cell_type":"code","source":["from pyspark.sql.types import *"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"ff021f33-743f-40b0-8258-41ba65c6b9d3"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["features = [\"EXT1\", \"EXT2\" ,\"EXT3\" ,\"EXT4\" ,\"EXT5\" ,\"EXT6\" ,\"EXT7\" ,\"EXT8\" ,\"EXT9\" ,\"EXT10\",\n            \"EST1\" ,\"EST2\" ,\"EST3\" ,\"EST4\" ,\"EST5\" ,\"EST6\" ,\"EST7\" ,\"EST8\" ,\"EST9\" ,\"EST10\",\n            \"AGR1\" ,\"AGR2\" ,\"AGR3\" ,\"AGR4\" ,\"AGR5\" ,\"AGR6\" ,\"AGR7\" ,\"AGR8\" ,\"AGR9\" ,\"AGR10\",\n            \"CSN1\" ,\"CSN2\" ,\"CSN3\" ,\"CSN4\" ,\"CSN5\" ,\"CSN6\" ,\"CSN7\" ,\"CSN8\" ,\"CSN9\" ,\"CSN10\",\n            \"OPN1\" ,\"OPN2\" ,\"OPN3\" ,\"OPN4\" ,\"OPN5\" ,\"OPN6\" ,\"OPN7\" ,\"OPN8\" ,\"OPN9\" ,\"OPN10\"\n           ]\n\n\nfor each_feature in df.columns:\n    if each_feature in features:\n            df = df.withColumn(each_feature, df[each_feature].cast(FloatType()))"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"8317e28f-d66d-4b5f-b10c-389e206495d9"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"markdown","source":["### 3.Exploring the data"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"6e72fa92-6055-4273-8146-23cb03b605d0"}}},{"cell_type":"code","source":["from pyspark.ml.stat import Correlation\nfrom pyspark.ml.feature import VectorAssembler\n\n# convert to vector column first\nvector_col = \"corr_features\"\nassembler = VectorAssembler(inputCols=df[features].columns, outputCol=vector_col)\ndf_vector = assembler.transform(df[features]).select(vector_col)\n\n# get correlation matrix\nmatrix = Correlation.corr(df_vector, vector_col)"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"35d5066f-fcf9-4855-b6fa-7ab2ef572c90"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["from matplotlib import pyplot as plt\n# import matplotlib as plt\nimport pandas as pd\nimport numpy as np\nmatrix =  matrix.collect()\na = matrix[0][0].toArray()\nplt.figure(figsize=(25, 25))\nplt.imshow(a, cmap=\"hot\", interpolation=\"nearest\")\nplt.colorbar()\nplt.xticks(ticks=range(50), labels=range(50))\nplt.xlabel(\"q\")\nplt.yticks(ticks=range(50), labels=range(50))\nplt.ylabel(\"p\")"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"485e5e42-421d-4f8e-b2cc-9991d564a2f5"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Out[7]: </div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Out[7]: </div>"]}},{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"/plots/199da95c-dfc4-4721-85c6-151feb962fd3.png","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"image","arguments":{}}},"output_type":"display_data","data":{"image/png":"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"}},{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Text(0, 0.5, &#39;p&#39;)</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Text(0, 0.5, &#39;p&#39;)</div>"]}}],"execution_count":0},{"cell_type":"markdown","source":["### 4.Preparing LabeledPoint for mllib"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"b2f0dc5e-9d49-4e47-9c7b-fc8e7a1917f3"}}},{"cell_type":"code","source":["categorical_feats = ['country']\nfrom pyspark.ml.feature import StringIndexer\n\ntd = df\nfor feat in categorical_feats:\n    stringIndexer = StringIndexer(inputCol=feat, outputCol=feat+\"_indexed\")\n    model = stringIndexer.fit(df)\n    df = model.transform(df)"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"8e7cf330-2cb9-401c-9ec3-8c3de91b08a5"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["features = [\"EXT1\", \"EXT2\" ,\"EXT3\" ,\"EXT4\" ,\"EXT5\" ,\"EXT6\" ,\"EXT7\" ,\"EXT8\" ,\"EXT9\" ,\"EXT10\",\n            \"EST1\" ,\"EST2\" ,\"EST3\" ,\"EST4\" ,\"EST5\" ,\"EST6\" ,\"EST7\" ,\"EST8\" ,\"EST9\" ,\"EST10\",\n            \"AGR1\" ,\"AGR2\" ,\"AGR3\" ,\"AGR4\" ,\"AGR5\" ,\"AGR6\" ,\"AGR7\" ,\"AGR8\" ,\"AGR9\" ,\"AGR10\",\n            \"CSN1\" ,\"CSN2\" ,\"CSN3\" ,\"CSN4\" ,\"CSN5\" ,\"CSN6\" ,\"CSN7\" ,\"CSN8\" ,\"CSN9\" ,\"CSN10\",\n            \"OPN1\" ,\"OPN2\" ,\"OPN3\" ,\"OPN4\" ,\"OPN5\" ,\"OPN6\" ,\"OPN7\" ,\"OPN8\" ,\"OPN9\" ,\"OPN10\", \"country_indexed\"\n           ]"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"7cb0663b-0435-4704-8303-7f2d21adafe9"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["new_df = df.select(features)"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"329a97ca-4618-45fa-af9d-fd3c164cbe83"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["from pyspark.mllib.regression import LabeledPoint\nno_0_df = new_df.rdd.filter(lambda row: row[0] > 0) # Make sure all labels are in 1 to 5\nlabelpointRDD_no_0 = no_0_df.map(lambda row:LabeledPoint(row[0], row[1:])) # convert the data into format of LabelPoint\nlabelpointRDD_no_0 = labelpointRDD_no_0.map(lambda x: LabeledPoint(x.label-1, x.features)) # convert label from 1-5 to 0-4\ntrain_no_0, test_no_0 = labelpointRDD_no_0.randomSplit([0.8, 0.2]) # Split training and test sets"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"d47a9c60-3abe-464e-bdbf-5fd5e14a7ba1"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"markdown","source":["### 5.Multiclass logistic regression"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"ffc5cfc4-beb0-4b1c-811e-4e2b2c4b9fd3"}}},{"cell_type":"code","source":["from pyspark.mllib.classification import LogisticRegressionWithLBFGS, LogisticRegressionModel\nfrom pyspark.mllib.regression import LabeledPoint\nmodel = LogisticRegressionWithLBFGS.train(train_no_0, numClasses=5)"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"466e010a-8fba-42e2-a723-418b8f446b3e"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["from pyspark.mllib.evaluation import MulticlassMetrics\npredictionAndLabels = test_no_0.map(lambda lp: (float(model.predict(lp.features)), lp.label))\nmetrics = MulticlassMetrics(predictionAndLabels)\nprecision = metrics.precision(1.0)\nrecall = metrics.recall(1.0)\nf1Score = metrics.fMeasure(1.0)\nprint(\"Summary Stats\")\nprint(\"Precision = %s\" % precision)\nprint(\"Recall = %s\" % recall)\nprint(\"F1 Score = %s\" % f1Score)\nprint(\"Accuracy = %s\" % metrics.accuracy)"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"9a1ecd6d-28e1-4928-8185-c40127fd54d2"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Summary Stats\nPrecision = 0.3739029800366477\nRecall = 0.09736068908364932\nF1 Score = 0.15449292687786412\nAccuracy = 0.46834784933724827\n</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Summary Stats\nPrecision = 0.3739029800366477\nRecall = 0.09736068908364932\nF1 Score = 0.15449292687786412\nAccuracy = 0.46834784933724827\n</div>"]}}],"execution_count":0},{"cell_type":"markdown","source":["### 6.Random Forest"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"48d027ce-3b28-4d5a-b471-90093474748b"}}},{"cell_type":"code","source":["from pyspark.mllib.tree import RandomForest, RandomForestModel\n\nmodel = RandomForest.trainClassifier(train_no_0, numClasses=5, categoricalFeaturesInfo={},\n                                     numTrees=5, featureSubsetStrategy=\"auto\",\n                                     impurity='gini', maxDepth=4, maxBins=32)"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"4cc943c2-9b06-4de9-9f86-06e1b0ab5f3b"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\"></div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\"></div>"]}}],"execution_count":0},{"cell_type":"code","source":["predictions = model.predict(test_no_0.map(lambda x: x.features))\nlabelsAndPredictions = test_no_0.map(lambda lp: lp.label).zip(predictions)\nacc = labelsAndPredictions.filter(lambda x: x[0] == x[1]).count() / float(test_no_0.count())\nprint(\"Model accuracy: %.3f%%\" % (acc * 100))"],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"7ebe75e8-952b-42e9-9ec1-2cc73a827f5c"}},"outputs":[{"output_type":"display_data","metadata":{"application/vnd.databricks.v1+output":{"datasetInfos":[],"data":"<div class=\"ansiout\">Model accuracy: 45.257%\n</div>","removedWidgets":[],"addedWidgets":{},"metadata":{},"type":"html","arguments":{}}},"output_type":"display_data","data":{"text/html":["<style scoped>\n  .ansiout {\n    display: block;\n    unicode-bidi: embed;\n    white-space: pre-wrap;\n    word-wrap: break-word;\n    word-break: break-all;\n    font-family: \"Source Code Pro\", \"Menlo\", monospace;;\n    font-size: 13px;\n    color: #555;\n    margin-left: 4px;\n    line-height: 19px;\n  }\n</style>\n<div class=\"ansiout\">Model accuracy: 45.257%\n</div>"]}}],"execution_count":0},{"cell_type":"code","source":[""],"metadata":{"application/vnd.databricks.v1+cell":{"title":"","showTitle":false,"inputWidgets":{},"nuid":"b5c49543-cb58-4f94-a0ca-32e4391f30ad"}},"outputs":[],"execution_count":0}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"mimetype":"text/x-python","name":"python","pygments_lexer":"ipython3","codemirror_mode":{"name":"ipython","version":3},"version":"3.8.5","nbconvert_exporter":"python","file_extension":".py"},"application/vnd.databricks.v1+notebook":{"notebookName":"trail","dashboards":[],"notebookMetadata":{"pythonIndentUnit":2},"language":"python","widgets":{},"notebookOrigID":2187261945615891}},"nbformat":4,"nbformat_minor":0}