Differential D23189 Diff 63821 src/backend/generalTest/HypothesisTest.cpp

Changeset View

Standalone View

View Options

src/backend/generalTest/HypothesisTest.cpp

Context not available.
32#include "backend/core/column/Column.h" 32​#include "backend/core/column/Column.h"
33#include "backend/lib/macros.h" 33​#include "backend/lib/macros.h"
34 34​
35#include <QVector> ​
36#include <QStandardItemModel> 35​#include <QStandardItemModel>
37#include <QLocale> ​
38#include <QLabel> 36​#include <QLabel>
37​#include <QTableView>
39#include <QVBoxLayout> 38​#include <QVBoxLayout>
40#include <QWidget> 39​#include <QCheckBox>
41#include <QtMath> ​
42#include <QQueue> ​
43 40​
44#include <KLocalizedString> 41​#include <KLocalizedString>
45 42​
Context not available.
68 m_tail = tail; 65​ m_tail = tail;
69} 66​}
70 67​
71void HypothesisTest::performTest(int test, bool categoricalVariable, bool equalVariance) { 68​void HypothesisTest::performTest(int test, bool categoricalVariable, bool equalVariance, bool calculateStats) {
72 m_pValue.clear(); 69​ m_pValue.clear();
73 m_statisticValue.clear(); 70​ m_statisticValue.clear();
74 m_statsTable = ""; 71​ m_statsTable = "";
Context not available.
77 m_resultLine[i]->setToolTip(QString()); 74​ m_resultLine[i]->setToolTip(QString());
78 } 75​ }
79 76​
77​ if (calculateStats)
78​ m_inputStatsTableModel->clear();
79​
80 switch (testSubtype(test)) { 80​ switch (testSubtype(test)) {
81 case TwoSampleIndependent: { 81​ case TwoSampleIndependent: {
82 m_currTestName = "<h2>" + i18n("Two Sample Independent Test") + "</h2>"; 82​ m_currTestName = "<h2>" + i18n("Two Sample Independent Test") + "</h2>";
83 performTwoSampleIndependentTest(testType(test), categoricalVariable, equalVariance); 83​ performTwoSampleIndependentTest(testType(test), categoricalVariable, equalVariance, calculateStats);
84 break; 84​ break;
85 } 85​ }
86 case TwoSamplePaired: 86​ case TwoSamplePaired:
Context not available.
121 emit changed(); 121​ emit changed();
122} 122​}
123 123​
124​void HypothesisTest::initInputStatsTable(int test, bool calculateStats) {
125​ m_inputStatsTableModel->clear();
126​
127​ if (!calculateStats) {
128​ switch (testSubtype(test)) {
129​ case TwoSampleIndependent: {
130​ m_inputStatsTableModel->setRowCount(2);
131​ m_inputStatsTableModel->setColumnCount(4);
132​ m_inputStatsTableModel->setHorizontalHeaderLabels(
133​ {i18n("N"), i18n("Sum"), i18n("Mean"), i18n("Standard Deviation")});
134​ break;
135​ }
136​ case TwoSamplePaired:
sgerlachUnsubmitted
Done

are these cases not implemented yet? If you only consider TwoSampleIndependent please use an if statement.

sgerlach: are these cases not implemented yet? If you only consider TwoSampleIndependent please use an if…
137​ case OneSample:
138​ case OneWay:
139​ case TwoWay:
140​ break;
141​ }
142​ }
143​
144​ emit changed();
145​}
146​
124QList<double>& HypothesisTest::statisticValue() { 147​QList<double>& HypothesisTest::statisticValue() {
125 return m_statisticValue; 148​ return m_statisticValue;
126} 149​}
Context not available.
135//TODO: backend of z test; 158​//TODO: backend of z test;
136//TODO: use https://www.gnu.org/software/gsl/doc/html/statistics.html for basic statistic calculations 159​//TODO: use https://www.gnu.org/software/gsl/doc/html/statistics.html for basic statistic calculations
137 160​
138 ​
139/**************************Two Sample Independent *************************************/ 161​/**************************Two Sample Independent *************************************/
162​// program is crashing on some input.
163​void HypothesisTest::performTwoSampleIndependentTest(int test, bool categoricalVariable, bool equalVariance, bool calculateStats) {
164​ int n[2];
165​ double sum[2], mean[2], std[2];
166​ QString col1Name;
167​ QString col2Name;
168​
169​ if (!calculateStats) {
170​ QString textValue;
171​ QDEBUG("m_inputStatsTable row and column count " << m_inputStatsTableModel->rowCount() << m_inputStatsTableModel->columnCount());
140 172​
141void HypothesisTest::performTwoSampleIndependentTest(int test, bool categoricalVariable, bool equalVariance) { 173​ for (int i = 0; i < 2; i++) {
174​ n[i] = m_inputStatsTableModel->data(m_inputStatsTableModel->index(i, 0)).toInt();
175​ sum[i] = m_inputStatsTableModel->data(m_inputStatsTableModel->index(i, 1)).toDouble();
176​ mean[i] = m_inputStatsTableModel->data(m_inputStatsTableModel->index(i, 2)).toDouble();
177​ std[i] = m_inputStatsTableModel->data(m_inputStatsTableModel->index(i, 3)).toDouble();
178​
179​ if (sum[i] == 0.0)
180​ sum[i] = mean[i] * n[i];
181​ if (mean[i] == 0.0 && n[i] > 0)
182​ mean[i] = sum[i] / n[i];
183​
184​ col1Name = "1";
sgerlachUnsubmitted
Done

this is set twice, right? If it's constant you can initialize the QStrings with it.

sgerlach: this is set twice, right? If it's constant you can initialize the QStrings with it.
devanshuagarwalAuthorUnsubmitted
Done

Not actually. It is set twice with different values based on the block.
So there is no use of initializing the two QStrings with some value.

devanshuagarwal: Not actually. It is set twice with different values based on the block. So there is no use of…
sgerlachUnsubmitted
Done

col1Name="1" and col2Name="2" are inside the for-loop and set twice with the same value or am i missing something?

sgerlach: col1Name="1" and col2Name="2" are inside the for-loop and set twice with the same value or am i…
devanshuagarwalAuthorUnsubmitted
Not Done

Yes, Sorry. I was error. I rectified it now. Thankyou.

devanshuagarwal: Yes, Sorry. I was error. I rectified it now. Thankyou.
185​ col2Name = "2";
186​ }
187​ } else {
142 if (m_columns.size() != 2) { 188​ if (m_columns.size() != 2) {
143 printError("Inappropriate number of m_columns selected"); 189​ printError("Inappropriate number of m_columns selected");
144 return; 190​ return;
145 } 191​ }
146 192​
147 int n[2]; 193​ col1Name = m_columns[0]->name();
148 double sum[2], mean[2], std[2]; 194​ col2Name = m_columns[1]->name();
149 ​
150 QString col1Name = m_columns[0]->name(); ​
151 QString col2Name = m_columns[1]->name(); ​
152 195​
153 if (!categoricalVariable && m_columns[0]->isNumeric()) { 196​ if (!categoricalVariable && m_columns[0]->isNumeric()) {
154 for (int i = 0; i < 2; i++) { 197​ for (int i = 0; i < 2; i++) {
Context not available.
201 col2Name = baseColName + " " + iter.key(); 244​ col2Name = baseColName + " " + iter.key();
202 } 245​ }
203 } 246​ }
247​ }
204 248​
205 QVariant rowMajor[] = {"", "N", "Sum", "Mean", "Std", 249​ QVariant rowMajor[] = {"", "N", "Sum", "Mean", "Std",
206 col1Name, n[0], sum[0], mean[0], std[0], 250​ col1Name, n[0], sum[0], mean[0], std[0],
Context not available.
211 255​
212 for (int i = 0; i < 2; i++) { 256​ for (int i = 0; i < 2; i++) {
213 if (n[i] == 0) { 257​ if (n[i] == 0) {
258​ if (calculateStats)
214 printError("At least two values should be there in every column"); 259​ printError("At least two values should be there in every column");
260​ else
261​ printError("In Statistic Table you filled above, value of N for every row should be > 0");
215 return; 262​ return;
216 } 263​ }
217 if (std[i] == 0.0) { 264​ if (std[i] == 0.0) {
Context not available.
250 printLine(9, "<b>Assumption:</b> UnEqual Variance b/w both population means"); 297​ printLine(9, "<b>Assumption:</b> UnEqual Variance b/w both population means");
251 } 298​ }
252 299​
300​ printLine(6, i18n("Degree of Freedom is %1", df), "green");
301​ printTooltip(6, i18n("Number of independent Pieces of information that went into calculating the estimate"));
302​
253 printLine(8, "<b>Assumption:</b> Both Populations approximately follow normal distribution"); 303​ printLine(8, "<b>Assumption:</b> Both Populations approximately follow normal distribution");
254 break; 304​ break;
255 } 305​ }
Context not available.
272 printTooltip(4, i18n("More is the |%1-value|, more safely we can reject the null hypothesis", testName)); 322​ printTooltip(4, i18n("More is the |%1-value|, more safely we can reject the null hypothesis", testName));
273 323​
274 printLine(5, i18n("P Value is %1 ", m_pValue[0]), "green"); 324​ printLine(5, i18n("P Value is %1 ", m_pValue[0]), "green");
275 ​
276 printLine(6, i18n("Degree of Freedom is %1", df), "green"); ​
277 printTooltip(6, i18n("Number of independent Pieces of information that went into calculating the estimate")); ​
278 ​
279 printTooltip(5, getPValueTooltip(m_pValue[0])); 325​ printTooltip(5, getPValueTooltip(m_pValue[0]));
326​
280 return; 327​ return;
281} 328​}
282 329​
Context not available.