Mathematics for Time Series

The goal of this library is to contain a bunch of common time-series analysis tools and indicators mostly used in technical analysis in finance.

List of contents

• Linear Trend Line
• Relative Strength Index
• Moving Average Convergence Divergence
• Bollinger Bands
• Ichimoku Cloud
• Fibonacci Retracement
• Fibonacci Fan
• Fibonacci Arc
• Fibonacci Time Zones

Build Instructions

You can use Maven or Gradle to add the library to your project.

Maven

<dependency>
    <groupId>com.github.ahmadmo</groupId>
    <artifactId>math-time-series</artifactId>
    <version>0.0.4</version>
</dependency>

Gradle

dependencies {
    compile 'com.github.ahmadmo:math-time-series:0.0.4'
}

Linear Trend Line

Imports

import ir.ahmadmo.math.timeseries.trend.*
import org.knowm.xchart.SwingWrapper
import org.knowm.xchart.XYChartBuilder
import org.knowm.xchart.style.Styler
import org.knowm.xchart.style.markers.SeriesMarkers
import kotlin.math.abs
import kotlin.math.min

You need to add the dependency for xchart to run the sample code:

<dependency>
    <groupId>org.knowm.xchart</groupId>
    <artifactId>xchart</artifactId>
    <version>3.6.5</version>
</dependency>

Sample Code

val trend = generateRandomTrend(size = 120)
val smoothingFactors = doubleArrayOf(0.01, 0.02, 0.05, 0.1, 0.15, 0.25)

val (resistance, support, average) = trend.linearTrendLines(smoothingFactors)

val resistanceData = DoubleArray(trend.size, resistance::value)
val supportData = DoubleArray(trend.size, support::value)
val averageData = DoubleArray(trend.size, average::value)

Notice that this is the user's responsibility to choose the rigth values for smoothingFactors from 0.0 to 1.0 to achieve better results. Choosing higher values results in smoother trends and consequently tighter trend lines. An empty smoothingFactors disables smoothing.

The function for generating a random trend (time-series):

fun generateRandomTrend(size: Int): Trend {
    val trend = Trend(size)
    var min = Double.MAX_VALUE
    for (x in trend.indices.drop(1)) {
        trend[x] = trend[x - 1] + Math.random() - 0.5
        min = min(min, trend[x])
    }
    val shift = abs(min(min, 0.0))
    return if (shift == 0.0) trend
    else Trend(size) { x -> trend[x] + shift }
}

Finally, to visualize the computed trend lines using xchart: (learn more)

val chart = XYChartBuilder()
    .width(1280).height(720)
    .theme(Styler.ChartTheme.Matlab)
    .title("Trend Lines")
    .xAxisTitle("Day").yAxisTitle("Price")
    .build()

chart.styler.isPlotGridLinesVisible = false
chart.styler.xAxisTickMarkSpacingHint = 100

val xData = DoubleArray(trend.size, Int::toDouble)

chart.addSeries("Trend", xData, trend)
chart.addSeries("Resistance", xData, resistanceData)
chart.addSeries("Support", xData, supportData)
chart.addSeries("Average", xData, averageData)

chart.seriesMap.values.forEach { it.marker = SeriesMarkers.NONE }

SwingWrapper(chart).displayChart()

Here is the complete sample code demonstrating how to draw linear trend lines for a time-series.

Sample Results

Relative Strength Index (RSI)

TODO

Moving Average Convergence Divergence (MACD)

TODO

Bollinger Bands

TODO

Ichimoku Cloud

TODO

Fibonacci Retracement

TODO

Fibonacci Fan

TODO

Fibonacci Arc

TODO

Fibonacci Time Zones

TODO