BenchmarkDP - Format technology lifecycle analysis - DOCUMENTS market segment

This set of files contains all results from analyzing DOCUMENTS market segment.<br><br>The fileset contains the following files:<br> <br> <i>experimentConfig.tsv</i> - A file containing TAB separated values. This file contains experiment configuration data. <br> <br> <i>market.tsv</i> - A file containing TAB separated values. This file contains definition of market elements. <br> <br> <i>marketStats.tsv </i>- A file containing TAB separated values. It contains statistics around data preparation step where the main aspect is conflict reduction. <br> <br> <i>cleanedData.tsv</i> - A file containing TAB separated values. This file contains the data for a specified market segment. The data is in the stage where the UK web archive Format Profile data has been filtered and all the conflicts have been resolved. <br> <br> <i>adoptionRates.tsv</i> - A file containing TAB separated values. This file extends cleanedData.tsv with the calculated adoption rates and smoothed adoption rates for each market element. Smoothed adoption rates are calculated as a three year moving average. <br> <br> <i>market elements.zip</i> - A zip file containing for each market element estimated models for all possible combinations of starting points. Parameter values are stored in allEstimates.tsv and calculated models (R objects) are in models folder. Furthermore, maximum 5 best models are selected (best.tsv) and their plots are in the graphs folder. <br> <br> <i>selected models.zip</i> - A zip file containing plots from selected models. Each market element is represented with the main plot and two plots showing residuals. <br> <br> <i>predictions.zip</i> - A zip file containing plots from selected models. In this case the prediction band (95%) is plotted instead of the confidence band. This zip contains also the prediction values for each market element in a file *-predictions.tsv where * is the name of the market element. Each file provides the actual value (if available) and the predicted value for a certain year (2009,2010,2011). <br>

本文件集涵盖针对文档市场细分（DOCUMENTS market segment）开展分析所得的全部结果。

本文件集包含如下文件：

1. 实验配置文件（experimentConfig.tsv）：采用制表符分隔值格式，存储实验配置相关数据。

2. 市场文件（market.tsv）：采用制表符分隔值格式，存储市场要素的定义信息。

3. 市场统计文件（marketStats.tsv）：采用制表符分隔值格式，存储围绕数据准备环节的统计信息，该环节的核心目标为冲突消解。

4. 清理后数据文件（cleanedData.tsv）：采用制表符分隔值格式，存储指定市场细分的数据集。该数据集已完成英国网络存档格式规范（UK web archive Format Profile）数据的过滤操作，且所有冲突均已完成消解。

5. 采用率文件（adoptionRates.tsv）：采用制表符分隔值格式，在cleanedData.tsv的基础上新增了各市场要素的计算采用率与平滑化采用率数据。平滑化采用率通过三年移动平均法计算得到。

6. 市场要素压缩包（market elements.zip）：内含各市场要素针对所有可能起始点组合的预估模型。参数值存储于allEstimates.tsv中，计算得到的模型（R语言对象）存放于models文件夹。此外，已筛选出至多5个最优模型（best.tsv），其对应可视化图表存放于graphs文件夹。

7. 选定模型压缩包（selected models.zip）：内含选定模型的可视化图表。每个市场要素对应一张主图表，以及两张展示残差的图表。

8. 预测结果压缩包（predictions.zip）：内含选定模型的可视化图表，此处将95%置信度的预测区间替代置信区间进行绘制。该压缩包同时包含各市场要素的预测值数据，存储于*-predictions.tsv文件中（其中*代表对应市场要素的名称）。每个文件均提供特定年份（2009、2010、2011）的实际值（若有）与预测值。