Ebffiledoc 3513
Ebffiledoc 3513
Ebffiledoc 3513
https://textbookfull.com/product/intelligence-science-and-big-
data-engineering-8th-international-conference-
iscide-2018-lanzhou-china-august-18-19-2018-revised-selected-
papers-yuxin-peng/
https://textbookfull.com/product/soft-computing-systems-second-
international-conference-icscs-2018-kollam-india-
april-19-20-2018-revised-selected-papers-ivan-zelinka/
Parallel Computational Technologies 12th International
Conference PCT 2018 Rostov on Don Russia April 2 6 2018
Revised Selected Papers Leonid Sokolinsky
https://textbookfull.com/product/parallel-computational-
technologies-12th-international-conference-pct-2018-rostov-on-
don-russia-april-2-6-2018-revised-selected-papers-leonid-
sokolinsky/
https://textbookfull.com/product/cloud-computing-and-
security-4th-international-conference-icccs-2018-haikou-china-
june-8-10-2018-revised-selected-papers-part-i-xingming-sun/
https://textbookfull.com/product/cloud-computing-and-
security-4th-international-conference-icccs-2018-haikou-china-
june-8-10-2018-revised-selected-papers-part-ii-xingming-sun/
https://textbookfull.com/product/cloud-computing-and-
security-4th-international-conference-icccs-2018-haikou-china-
june-8-10-2018-revised-selected-papers-part-vi-xingming-sun/
123
Communications
in Computer and Information Science 872
Commenced Publication in 2007
Founding and Former Series Editors:
Phoebe Chen, Alfredo Cuzzocrea, Xiaoyong Du, Orhun Kara, Ting Liu,
Dominik Ślęzak, and Xiaokang Yang
Editorial Board
Simone Diniz Junqueira Barbosa
Pontifical Catholic University of Rio de Janeiro (PUC-Rio),
Rio de Janeiro, Brazil
Joaquim Filipe
Polytechnic Institute of Setúbal, Setúbal, Portugal
Igor Kotenko
St. Petersburg Institute for Informatics and Automation of the Russian
Academy of Sciences, St. Petersburg, Russia
Krishna M. Sivalingam
Indian Institute of Technology Madras, Chennai, India
Takashi Washio
Osaka University, Osaka, Japan
Junsong Yuan
University at Buffalo, The State University of New York, Buffalo, USA
Lizhu Zhou
Tsinghua University, Beijing, China
More information about this series at http://www.springer.com/series/7899
Youness Tabii Mohamed Lazaar
•
123
Editors
Youness Tabii Mohammed Al Achhab
Abdelmalek Essaâdi University Abdelmalek Essaâdi University
Tétouan Tétouan
Morocco Morocco
Mohamed Lazaar Nourddine Enneya
Abdelmalek Essaâdi University Université Ibn-Tofail
Tétouan Tétouan
Morocco Morocco
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
We are happy to present you this book, Big Data, Cloud and Applications, which is a
collection of papers that were presented at the 3rd International Conference on Big
Data Cloud and Applications, BDCA 2018. The conference took place on April 04–05,
2018, in Kenitra, Morocco.
The book consisted of nine chapters, which correspond to the four major areas that
are covered during the conference, namely, Big Data, Cloud Computing, Maching
Learning, Deep Learning, Data Analysis, Neural Networks, Information System and
Social Media, Natural Language Processing, Image Processing and Applications.
Every year BDCA attracted researchers from all over the world, and this year was
not an exception – we received 99 submissions from 12 countries. More importantly,
there were participants from many countries, which indicates that the conference is
truly gaining more and more international recognition as it brought together a vast
number of specialists who represented the aforementioned fields and share information
about their newest projects. Since we strived to make the conference presentations and
proceedings of the highest quality possible, we only accepted papers that presented the
results of various investigations directed to the discovery of new scientific knowledge
in the area of Big Data, Cloud Computing and their applications. Hence, only 45 papers
were accepted for publishing (i.e., 45% acceptance rate). All the papers were reviewed
and selected by the Program Committee, which comprised 96 reviewers from over 58
academic institutions. As usual, each submission was reviewed following a double
process by at least two reviewers. When necessary, some of the papers were reviewed
by three or four reviewers. Our deepest thanks and appreciation go to all the reviewers
for devoting their precious time to produce truly through reviews and feedback to the
authors.
The 3rd International Conference on Big Data, Cloud and Applications (BDCA 2018)
was organized by Abdelmalek Essaadi University and IbnTofail University and was in
Kenitra, Morocco (April 04–05, 2018).
General Chairs
Youness Tabii National School of Applied Sciences (ENSA), Tetouan,
Morocco
Nourddine Enneya Faculty of Sciences, Kenitra, Morocco
Program Committee
Hamid R. Arabnia University of Georgia, USA
Abdelkaher Ait Ibn Zohr University, Morocco
Abdelouahad
Noura Aknin FS, Abdelmalek Essaadi University, Morocco
Adel Alimi REGIM, Sfax University, Tunisia
Mohammed Al Achhab ENSA, Abdelmalek Essaadi University, Morocco
Naoual Attaoui FS, Abdelmalek Essaadi University, Morocco
Abderrahim Azouani Mohammed 1st University, Morocco
Jenny Benois-Pineau Bordeaux University, France
Abdellah Abouabdellah ENSA, Ibn Tofail University, Morocco
Amel Benazza Supcom Carthage University, Tunisia
VIII Organization
Big Data
Cloud Computing
Machine Learning
Deep Learning
Data Analysis
Splitting Method for Decision Tree Based on Similarity with Mixed Fuzzy
Categorical and Numeric Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Houda Zaim, Mohammed Ramdani, and Adil Haddi
Contents XIII
Neural Networks
An Enhanced MSER Based Method for Detecting Text in License Plates. . . . 464
Mohamed Admi, Sanaa El Fkihi, and Rdouan Faizi
1 Introduction
The learning is a long life process which takes place everywhere; it is divided in two
categories [1] formal and non-formal or informal. Formal learning is often validated by
official certifications; education occurs in structured environments such as schools and
universities and is supervised by teachers. Knowledge and skills acquired outside the
formal setting enable an informal learning. In today’s world, communication between
people occurs often through the use of social media platforms, wikis, micro-blogs which
become the main channels for conveying and sharing information in a quickly manner.
Communities and groups have been built around common points of interest. With
advances in Web2 technologies, the user of social network platforms once authenticated,
he can freely have several roles, read other people’s posts, write messages, insert media
and documents, search people and trend topics. Although social networks are considered
as entertainment spaces, several universities are attracted by the insertion of informal
learning via social networks like Twitter in their academic development [2]. In fact, in
this new age of data and computing, many individuals, students in higher education or
professionals have resorted to informal means to educate themselves and upgrade their
skills for example in cutting edge of tools in information technologies by working on-
line short courses and workshops. Informal learning through social media leads to
empowerment and self-efficacy while saving time and money in the learning process
and increase visibility in society.
Social network Analytics [3] is a set of methods and technologies that allow
collecting a large datasets from social network platforms sources, transform them in a
way that they become available and ready to be consumed by analysts. Text mining,
Natural language processing, classification and clustering algorithms are used to extract
the hidden insights in order to improve the best knowing of the user’s experiences. New
open source technologies like Apache Hadoop [4] and Spark [5] allow building infra‐
structures which aimed to manage massive datasets by distributing storage and
computing across clusters of low cost machines, they handle and combine both struc‐
tured and unstructured data that come from internal and external data sources.
Depending on data production, data processing tasks is divided into two groups:
– Batch Processing: data are collected in big batches over period of time, it is stored
in distributed file system, then processing and analysis jobs are applied at once, and
batch results are generated.
– Streaming Processing: data come in continuous way; processing and analysis jobs
are applied in near real time or in small time. In this work we use Apache Spark as
data processing engine, it is a distributed framework developed in Scala programming
language and works as a Java Virtual Machine, Spark is designed for fast scalable
in-memory computing and relies on Hadoop to run in cluster mode and use HDFS
[6] storage, it comes with a high level programming model that hides the partitioning
of dataset in memory of cluster, using a novel data structure called Resilient Distrib‐
uted Dataset RDD [7] which is an immutable distributed collection of objects parti‐
tioned across different nodes of the cluster. RDD data-sharing abstraction allows to
use wide range of APIs provided by Spark: Spark SQL, Spark Streaming, MLlib
(Machine Learning library), and GraphX (graph processing). Apache spark is suited
to perform analytics that need iterative operations. It allows to process data directly,
comparing to Map Reduce [8] programs which need several access to disk to retrieve
intermediate result. Since twitter data are generated at high speed and in a complex
structure, we implement a hybrid architecture which provides a faster ETL based on
data pipeline that ensures the data collection and processing in a unified and distrib‐
uted environment. We have conducted an experiment on twitters data filtered by
keywords associated with 6 topics of big data technologies and data science which
are of hot interests to developer and industrial communities.
In this paper we describe the necessary steps to carry out an exploratory analysis and
the extraction of the top group of connected hashtags.
Informal Learning in Twitter 5
The rest of the paper is structured as follows. Section 2 discusses related work.
Section 3 describes the Architecture of Data analysis workflow, Sect. 4 discusses the
experiment and finally, Sect. 5 concludes the paper.
2 Related Work
Social network analysis is an emerging research field which aims to better understand
how people seek and share information in social network platforms. Bonchi et al. in [9]
provided an overview of what we consider to be key problems and techniques in social
network analysis from a business applications perspective. The authors described each
area of research in the context of a specific business processes classification framework
(The APQC process classification framework), and then focused on several areas, giving
an overview of the main problems and describing state-of-the-art approaches.
The explosion of the use of micro blogs by students offers opportunities to exploit
this new communication channel in process-oriented learning. In paper [10], the authors
proposed a platform which uses Twitter news in Education known as NIE in order to
provide the latest news classified on various topics then enable discussion and debate
groups. They implemented a prototype system which uses Twitter as source to the hot
news and trends. For classification topics, each news tweet is cleaned and mapped into
its words. The Naïve Bayes classifier is used to achieve the classification based on
predefined number of keywords which correspond to the selected topics. The platform
offers to learners a News Visualizer using treemap to facilitate the learner’s query which
is based on period, keywords, and desired topic. Cosine similarity method, Based on
user document similarity and hierarchical agglomerative clustering is used to study the
learners’ preferences.
Aramo-Immonen et al. [11] employ Twitter data to study interactions between
members of community of managers attending a conference. Data are retrieved two
weeks before the conference. The process of data-driven visual network analytics and
the Ostinato [12] process model are provided to extract insights into the informal
learning of community managers. Quantitative and qualitative analyses of Twitter data
are produced like analysis of the top hash tags over time before the conference and the
network of hash tag co-occurrences.
In paper [13], the authors developed a workflow that consists to integrate both qual‐
itative analysis and large-scale data mining techniques. They focused on engineering
students’ Twitter posts to understand issues and problems in their educational experi‐
ences. The authors conducted a qualitative analysis on samples taken from about 25,000
tweets related to engineering students’ college life. They found engineering students
encounter problems such as heavy study load, lack of social engagement, and sleep
deprivation. A multi-label classification algorithm is implemented to classify tweets
reflecting students’ problems.
The majority of tweets do not contain the geographical location through exact GPS
coordinates (latitude and longitude). The authors attempt in [14] to identify a location
of the tweets. They employ twitter data to fit a Naive bayes model in order to classify a
tweets based on features as users’ timezone, the user’s language, and the parsed users’
6 A. Chaffai et al.
location. The classifier with an accuracy of 82% was achieved and performs well on
active Twitter countries such as the Netherlands and United Kingdom.
An analysis of errors made by the classifier shows that mistakes were made due to
limited information and shared properties between countries such as shared timezone.
A feature analysis was performed in order to see the effect of different features. The
features timezone and parsed user location were the most informative features.
Twitter has become a largest social space in the world where 330 million monthly active
users, discuss several topics and publish 500 million tweets per day. This data source
offers tremendous opportunities to analyze social trends for multiple purposes. Twitter
offers two types of APIs, Rest API and streaming APIs (for developers in real time) that
allow different clients applications written in different languages [15] to consume the
tweets. For example, in case of Java and Scala, Twitter4J is an open source Java library
used for interfacing with Twitter’s Application Programming Interfaces (APIs). Tweets
data come in non-structured nature, they are encoded using Java Script Object Notation
(JSON) based on key-value pairs. Each tweet has an author (user), a message, a unique
ID, a timestamp of when it was created, and geo metadata often turned off by users. Each
User has a Twitter name, an ID, a number of followers. Tweet contains ‘entity’ objects,
which are arrays of contents such as hashtags, mentions, media, and links.
A typical SNA workflow consists of several interacting phases which are:
• Data collection
• Data preparation
• Data analysis
• Insights.
The different topics discussed in the context of informal learning and social learning
in twitter are very varied, in this paper we propose a flexible data system (see Fig. 1)
capable to receive data from different topics through multiple agents, each agent inter‐
cepts the stream data in real time based on keywords related to a given topic, Apache
Flume [16] is used in the data collection layer. We are faced with a case where there
will be several flume-agents, so we need a strategy to categorize the message, for this
we use Apache Kafka [17] as an efficient publish-subscribe messaging system to separate
the incoming data in topics and keep them in scalable and fault-tolerant way. In the rest
of data pipeline, we use Spark streaming to consume, parse the incoming data in real
time and store them in HDFS storage. Analysis tasks to extract insights can be performed
by using Spark SQL and Spark ML.
Informal Learning in Twitter 7
4 Experiment
Due to strong competition between organizations for integrating data into decision
making, hiring opportunities for data specialists and data infrastructure specialists are
much greater than those of other profiles. We will study this trend in the twitter social
network as a case study, to try to extract useful information about users who are inter‐
ested in acquiring new knowledge or who share their experiences in the field of big data.
We employ data from twitter that is filtered based on the following keywords: “bigdata”,
“datascience”, “machineLearning”, “hadoop”, “spark”, “analytics”.
We deployed a small local cluster for Hadoop and Spark on 11 nodes running Ubuntu
16.04 LTS and interconnected via one switch of 1 Gb/s. The Hadoop cluster is built
using Hadoop version 2.7.3. The Spark cluster is built using Spark version 2.0.0. One
machine is designed as Master for both Spark and Hadoop, the others nodes are both
the Hadoop slaves and Spark workers. The following configuration is the same for all
nodes: Intel(R) Core(TM) i5-3470 CPU 3.20 GHz(4CPUs), 1 Gb/s network connection,
300 GB hard disk, 8 GB Memory.
8 A. Chaffai et al.
4.3 Methodology
Data Ingestion
Retrieving data from the Twitter API requires credentials that can be obtained from
https://apps.twitter.com/, we register our application as a twitter app, then the authori‐
zation parameters are generated as follows: Consumer Key (API Key), Consumer Secret
(API Secret), Access Token and Access Token Secret.
Apache Flume is used to collect tweets data in JSON format from the source and
move it to Kafka in plaintext. As defined on its site [18], “Flume is a distributed and
available service for efficiently collecting, aggregating, and moving large amounts of
log data. It has a simple and flexible architecture based on streaming data flows.” The
main components of flume data pipeline (see Fig. 2) are source, channel, and sinks.
Flume agent is a JVM daemon responsible to manage the data flow. The source contin‐
uously retrieves tweets data in JSON format based on several keywords from the Twitter.
The channel act as a passive storage, it maintains the event data until a next hop which
is a Kafka cluster.
Data Processing
This phase consists to ingest data from Kafka topic for live processing in Apache Spark.
Since Spark is a batch processing, we use Spark streaming to retrieve continuously the
messages accumulated in Kafka topic. Spark streaming receives the input stream and
divides it in a series of mini batches corresponding to input periods equal to batch
interval, it creates a DStream (see Fig. 5) which is as a sequence of RDDs that can be
processed in Spark core as a static data.
Any streaming application needs a streaming context which is an entry point to the
Spark cluster resources.
We create our application in Scala that involves the following steps:
(1) To interact with kafka cluster, we connect spark streaming adopting the direct
approach using the DirectStream method in order to deploy a customized receiver
(see Fig. 6) which requires the subscription to bigdata_tweets topic created above.
10 A. Chaffai et al.
(2) Once the stream is created we convert it to JSON format (see Fig. 7), in order to
extract and process the interested fields in future analysis tasks. We store the stream
data in HDFS in JSON Format.
Insights
Exploratory Analysis
We collected 20058 tweets, that we stored in HDFS in JSON format, then we converted
them to DataFrame in a structured format appropriate to be queried. We create a table
by selecting the entities and fields in interest like text, hash tags, urls, place, user.lang
in order to extract insights using Spark SQL. Thus, we deduced that the tweets contain
several links to a diversified resources for informal learning which can adapt to all styles
of learning in the form of links to external pages, free tutorial and courses (see Table 1).
We have noticed the presence of several companies specialized in the eLearning industry
which publish their offers and course promotions to attract users interested in big data
technologies and data science.
We found 9214 distinct users, although geo-location is disabled in the majority of
tweets [14], but we can extract their origin from the time zone, and native languages,
we found that 80% of users are Americans.
4264 distinct hashtags found in tweets data, we extract the top 10 most popular
hashtags (see Fig. 8) with respectively the number of occurrences in all tweets.
Informal Learning in Twitter 11
Fig. 9. Sample of DataFrame created from raw data containing tweet identifier, user and hashtags.
In graph theory [19], graph is a data structure, conceptually described by a pair (S,
A) where S is a finite set of nodes called vertices or vertex and A is a finite multi-set of
ordered pairs of vertices called edges, an edge connects two vertices in a graph.
In real life applications, everything is interconnected, Graphs are mostly used to
represent the networks and model the relations between nodes, like routers, airports,
paths in cities, users in social networks.
A graph can be:
• Directed: the edges have a direction from the vertex source to the vertex destination
• Undirected: the edges have no direction.
• Directed multigraph: a pair of vertices is linked by one two or more edges, it describes
a multiple relationships. The edges share the same source and destination.
• Property Graphs: is a directed multigraph where vertex and edges have proprieties.
12 A. Chaffai et al.
A tweet can contain 0 to multiple hashtags, each hashtag represents a topic of discus‐
sion, the presence of multiple hashtags increase the engagement of the users and the
value of the publication. Using Scala, we implement a graph analytics pipeline with
Spark Graph X in order to convert the DataFrame (as shown in Fig. 9) to a graph and
find the top connected hashtags.
Building a graph with Graph X requires two arguments: RDD of Vertices and RDD
of edges, which can be instantiated based on two specialized RDD implementations:
– The VertexRDD[VD] is a parameterized class, it’s defined as RDD[(VertexId, VD)],
VertexId is a vertex identifier, it is an instance of Long, VD is the vertex attribute or
property it can be a user type defined or other type of data information that are related
to vertex.
– The EdgeRDD[ED] is a parameterized class which is an implementation of
RDD[Edge[ED]], an instance of Edge represents VertexId source, VertexId destina‐
tion, and the attribute of the property of the edge.
We build the structure of vertices from the hashtag name, for each hashtag we create
a unique identifier (VertexId) in 64 bit by using the MurmurHash3 library [20], the vertex
propriety takes the string value of the hashtag name.
For the edge which is the link between two nodes, a pair of hashtags is generated by
using the combinations function, since we have no information about the relationship
between hashtags except their presence in the same tweet we opt to use the Twitter
username as propriety of the edge. A triplet represents an edge with two connected
vertices. We employed data with the hashtags entities having a size greater or equal to
2 to avoid the appearance of isolated nodes in our graph. We present as follow (see
Fig. 10) the steps to generate the structures of vertices and edges:
From a pair of RDD vertices and edges, we create an instance of Graph class to
generate a graph data structure as follows: val graph = Graph(vertices, edges) (Fig. 11).
Informal Learning in Twitter 13
Fig. 11. Sample of graph vertices, graph edges and graph triplets.
Fig. 13. Sample of hashtags that belong to the top group of connected component
14 A. Chaffai et al.
5 Conclusion
In this paper we propose a social network analysis system designed around a Twitter
API source. This system is in the form of a real time data pipeline capable to capture
events which are the tweets related to informal learning and categorize them in topics
in order to extract valuable information. We combine Apache Flume and Kafka to build
the data ingestion layer which is responsible to retrieve live data. Apache Kafka cluster
is used for categorizing the data that transit. To process data in real time we use Spark
Streaming library. HDFS is used as a persistence layer.
This work is based on a real experience where we have collected a dataset of 20058
tweets, then we accomplished some steps to achieve the data pipeline analysis, and
finally we extracted the top group of connected hashtags using Spark Graph X API.
During this work we have identified new directions concerning the eLearning. The
first is to study the use of social network platforms by Moroccan students for informal
learning purposes, and the second is to study how to integrate social networks channels
in formal learning settings like eLearning platforms.
References
1. Cameron, R., Harrison, J.L.: The interrelatedness of formal, non-formal and informal
learning: evidence from labour market program participants. Aust. J. Adult Learn. 52(2), 277–
309 (2012)
2. McPherson, M., Budge, K., Lemon, N.: New practices in doing academic development:
Twitter as an informal learning space. Int. J. Acad. Dev. 20(2), 126–136 (2015)
3. Wadhwa, P., Bhatia, M.P.S.: Social networks analysis: trends, techniques and future
prospects. In: Fourth International Conference on Advances in Recent Technologies in
Communication and Computing (ARTCom 2012), Bangalore, India, pp. 1–6 (2012)
4. White, T.: Hadoop: The Definitive Guide. O’Reilly Media, Inc., Newton (2012)
5. Zaharia, M., Chowdhury, M., Franklin, M.J., Shenker, S., Stoica, I.: Spark: cluster computing
with working sets. In: Proceedings of the 2nd USENIX Conference on Hot Topics in Cloud
Computing (2010)
6. Ghemawat, S., et al.: The Google File System. ACM SIGOPS Operating Systems Review
(2013)
7. Zaharia, M., et al.: Resilient distributed datasets: a fault-tolerant abstraction for in-memory
cluster computing. In: Proceedings of 9th USENIX Conference on Networked Systems
Design and Implementation, p. 2 (2012)
8. Dean, J., Ghemawat, S.: MapReduce: simplified data processing on large clusters. Commun.
ACM 51(1), 107–113 (2008)
9. Bonchi, F., Castillo, C., Gionis, A., Jaimes, A.: Social network analysis and mining for
business applications. ACM Trans. Intell. Syst. Technol. (TIST) Arch. 2(3), 37 (2011). Article
22
10. Kim, Y., Hwang, E., Rho, S.: Twitter news-in-education platform for social collaborative and
flipped learning. J. Supercomput. Springer, 1–19 (2016). https://doi.org/10.1007/
s11227-016-1776-x
11. Aramo-Immonen, H., Kärkkäinen, H., Jussila, J.J., Joel-Edgar, S., Huhtamäki, J.: Visualizing
informal learning behavior from conference participants’ Twitter data with the Ostinato
model. J. Comput. Hum. Behav. Arch. 55(PA), 584–595 (2016)
Informal Learning in Twitter 15
12. Huhtamäki, J., Russell, M.G., Rubens, N., Still, K.: Ostinato: the exploration-automation cycle of
user-centric, process-automated data-driven visual network analytics. In: Matei, S., Russell, M.,
Bertino, E. (eds.) Transparency in Social Media, pp. 197–222. Cham, Computational Social
Sciences, Springer (2015). https://doi.org/10.1007/978-3-319-18552-1_11
13. Chen, X., Vorvoreanu, M., Madhavan, K.: Mining social media data for understanding
students’ learning experiences. IEEE Trans. Learn. Technol. 7(3), 246–259 (2014)
14. Chandra, S., Khan, L., Muhaya, F.B.: Estimating Twitter user location using social
interactions–a content based approach. In: IEEE Third International Conference on Privacy,
Security, Risk and Trust and 2011 IEEE Third International Conference on Social Computing,
Boston, MA, pp. 838–843 (2011)
15. Twitter libraries homepage. https://developer.twitter.com/en/docs/developer-utilities/
twitter-libraries. Accessed 24 Feb 2018
16. Shreedharan, H.: Using Flume. O’Reilly Media, Inc., Sebastopol (2014)
17. Vohra, D.: Apache kafka. In: Practical Hadoop Ecosystem. Apress, Berkeley, CA Apache
(2016)
18. Apache Flume homepage. https://flume.apache.org/. Accessed 24 Feb 2018
19. Bondy, J.A., Murty, U.S.R.: Graph Theory with Applications. American Elsevier Publishing
Company, New York (1976)
20. MurmurHash3 documentation. https://www.scala-lang.org/files/archive/api/2.11.0-M4/
index.html#scala.util.hashing.MurmurHash3$. Accessed 24 Feb 2018
A MapReduce-Based Adjoint Method to
Predict the Levenson Self Report
Psychopathy Scale Value
1 Introduction
The Dutch goodwife worked hard from early morn till sunset. She
worked in restricted ways, she had few recreations and pleasures
and altogether little variety in her life; but she possessed what
doubtless proved to her in that day, as it would to any woman in this
day, a source of just satisfaction, a soothing to the spirit, a staying of
melancholy, a moral support second only to the solace of religion,—
namely, a large quantity of very good clothes, which were
substantial, cheerful, and suitable, if not elegant.
The Dutch never dressed “in a plaine habbit according to the
maner of a poore wildernesse people,” as the Connecticut colonists
wrote of themselves to Charles II.; nor were they weary wanderers in
a wilderness as were Connecticut folk.
I have not found among the statutes of New Netherland any
sumptuary laws such as were passed in Connecticut,
Massachusetts, and Virginia, to restrain and attempt to prohibit
luxury and extravagance in dress. Nor have I discovered in the court-
records any evidences of magisterial reproof of finery; there is, on
the contrary, much indirect proof of encouragement to “dress orderly
and well according to the fashion and the time.” Of course the Dutch
had no Puritanical dread of over-rich garments; and we must also
never forget New Netherland was not under the control of a
government nor of a religious band, but of a trading-company.
The ordinary dress of the fair dames and damsels of New
Amsterdam has been vividly described by Diedrich Knickerbocker;
and even with the additional light upon their wardrobe thrown by the
lists contained in colonial inventories, I still think his description of
their every-day dress exceedingly good for one given by a man. He
writes:
“Their hair, untortured by the abominations of art, was
scrupulously pomatumed back from their foreheads with a
candle, and covered with a little cap of quilted calico, which
fitted exactly to their heads. Their petticoats of linsey-woolsey
were striped with a variety of gorgeous dyes, though I must
confess those gallant garments were rather short, scarce
reaching below the knee; but then they made up in the
number, which generally equalled that of the gentlemen’s
small-clothes; and what is still more praiseworthy, they were
all of their own manufacture,—of which circumstance, as may
well be supposed, they were not a little vain.
“Those were the honest days, in which every woman
stayed at home, read the Bible, and wore pockets,—ay, and
that, too, of a goodly size, fashioned with patchwork into
many curious devices, and ostentatiously worn on the
outside. These, in fact, were convenient receptacles where all
good housewives carefully stored away such things as they
wished to have at hand; by which means they often came to
be incredibly crammed.
“Besides these notable pockets, they likewise wore scissors
and pincushions suspended from their girdles by red ribbons,
or, among the more opulent and showy classes, by a brass
and even silver chains, indubitable tokens of thrifty
housewives and industrious spinsters. I cannot say much in
vindication of the shortness of the petticoats; it doubtless was
introduced for the purpose of giving the stockings a chance to
be seen, which were generally of blue worsted, with
magnificent red clocks; or perhaps to display a well-turned
ankle and a neat though serviceable foot, set off by a high-
heeled leathern shoe, with a large and splendid silver buckle.
“There was a secret charm in those petticoats, which no
doubt entered into the consideration of the prudent gallants.
The wardrobe of a lady was in those days her only fortune;
and she who had a good stock of petticoats and stockings
was as absolutely an heiress as is a Kamtschatka damsel
with a store of bear-skins, or a Lapland belle with plenty of
reindeer.”
A Boston lady, Madam Knights, visiting New York in 1704, wrote:
—
“The English go very fashionable in their dress. But the
Dutch, especially the middling sort, differ from our women, in
their habitt go loose, wear French muches wch are like a
Capp and head-band in one, leaving their ears bare, which
are sett out with jewells of a large size and many in number;
and their fingers hoop’t with rings, some with large stones in
them of many Coullers, as were their pendants in their ears,
which you should see very old women wear as well as
Young.”
This really gives a very good picture of the vrouws; “loose in their
habit,” wearing sacques and loose gowns, not laced in with pointed
waists as were the English and Boston women; with the ornamental
head-dress, and the gay display of stoned earrings and rings, which
was also not the usual wear of New England women, who generally
owned only a few funeral rings.
In the inventories of personal estates contained in the Surrogate’s
Court we find details of the wardrobe; but as I have enumerated and
defined all the different articles at some length in my book, “Costume
of Colonial Times,” I will not repeat the definitions here; but it should
be remembered that in the enumeration of the articles of clothing,
many stuffs and materials of simple names were often of
exceedingly good and even rich quality. From those inventories we
have proof that all Dutch women had plenty of clothes; while the
wives of the burgomasters, the opulent merchants, and those in
authority, had rich clothes. I have given in full in my book a list of the
clothing of a wealthy New York dame, Madam De Lange; but I wish
to refer to it again as an example of a really beautiful wardrobe. In it
were twelve petticoats of varying elegance, some worth two pounds
fifteen shillings each, which would be more than fifty dollars to-day.
They were of silk lined with silk, striped stuff, scarlet cloth, and ash-
gray cloth. Some were trimmed with gold lace. With those petticoats
were worn samares and samares-a-potoso, six in number, which
were evidently jackets or fancy bodies; these were of calico, crape,
“tartanel,” and silk. One trimmed with lace was worth three pounds.
Waistcoats and bodies also appear; also fancy sleeves. Love-hoods
of silk and cornet-caps with lace make a pretty head-gear to
complete this costume, with which was worn the reim or silver girdle
with hanging purse, and also with a handsome number of diamond,
amber, and white coral jewels.
The colors in the Dutch gowns were almost uniformly gay,—in
keen contrast to the sad-colored garments of New England. Madam
Cornelia de Vos in a green cloth petticoat, a red and blue
“Haarlamer” waistcoat, a pair of red and yellow sleeves, and a purple
“Pooyse” apron was a blooming flower-bed of color.
The dress of Vrouentje Ides Stoffelsen, a very capable
Dutchwoman who went to Bergen Point to live, varied a little from
that of these town dames. Petticoats she had, and waistcoats,
bodies and sleeves; but there was also homelier attire,—purple and
blue aprons, four pairs of pattens, a fur cap instead of love-hoods,
and twenty-three caps. She wore the simpler and more universal
head-gear,—a close linen or calico cap.
The head covering was of considerable importance in New
Amsterdam, as it was in Holland as well as in England at that date.
We find that it was also costly. In 1665 Mistress Piertje Jans sold a
fine “little ornamental headdress” for fifty-five guilders to the young
daughter of Evert Duyckinck. But it seems that Missy bought this
“genteel head-clothes” without the knowledge or permission of her
parents, and on its arrival at the Duyckinck home Vrouw Duyckinck
promptly sent back the emblem of extravagance and disobedience.
Summoned to court by the incensed milliner who wished no rejected
head-dresses on her hands, and who claimed that the transaction
was from the beginning with full cognizance of the parents, Father
Duyckinck pronounced the milliner’s bill extortionate; and
furthermore said gloomily, with a familiar nineteenth-century
phraseology of New York fathers, that “this was no time to be buying
and wearing costly head-dresses.” But the court decided in the
milliner’s favor.
It is to be deplored that we have no fashion-plates of past
centuries to show to us in exact presentment the varying modes
worn by New York dames from year to year; that method of fashion-
conveying has been adopted but a century. The modes in olden days
travelled from country to country, from town to town, in the form of
dolls or “babies,” as they were called, wearing miniature model
costumes. These dolls were dressed by cutters and tailors in Paris or
London, and with various tiny modish garments were sent out on
their important mission across the water. In Venice a doll attired in
the last fashions—the toilette of the year—was for centuries
exhibited on each Ascension Day at the “Merceria” for the edification
of noble Venetian dames, who eagerly flocked to the attractive sight.
Not less eagerly did American dames flock to provincial mantua-
makers and milliners to see the London-dressed babies with their
miniature garments. Even in this century, fashions were brought to
New York and Philadelphia and Albany through “milliners’ boxes”
containing dressed dolls. Mrs. Vanderbilt tells of one much admired
fashion-doll of her youth who had a treasured old age as a juvenile
goddess.
A leading man of New Amsterdam, a burgomaster, had at the time
of his death, near the end of Dutch rule, this plentiful number of
substantial garments: a cloth coat with silver buttons, a stuff coat,
cloth breeches, a cloth coat with gimp buttons, a black cloth coat, a
silk coat, breeches and doublet, a silver cloth breeches and doublet,
a velvet waistcoat with silver lace, a buff coat with silk sleeves, three
“gross-green” cloaks, several old suits of clothes, linen, hosiery,
silver-buckled shoes, an ivory-headed cane, and a hat. One hat may
seem very little with so many other garments; but the real beaver
hats of those days were so substantial, so well-made, so truly worthy
an article of attire, that they could be constantly worn and yet last for
years. They were costly; some were worth several pounds apiece.
Gayer masculine garments are told of in other inventories: green
silk breeches flowered with silver and gold, silver gauze breeches,
yellow fringed gloves, lacquered hats, laced shirts and neck-cloths,
and (towards the end of the century, and nearly through the
eighteenth century) a vast variety of wigs. For over a hundred years
these unnatural abominations, which bore no pretence of resembling
the human hair, often in grotesque, clumsy, cumbersome shapes,
bearing equally fantastic names, and made of various indifferent and
coarse materials, loaded the heads and lightened the pockets of our
ancestors. I am glad to note that they were taxed by the government
of the province of New York. The barber and wig-maker soon
became a very important personage in a community so given over to
costly modes of dressing the head. Advertisements in the
newspapers show the various kinds of wigs worn in the middle of the
eighteenth century. From the “New York Gazette” of May 9, 1737, we
learn of a thief’s stealing “one gray Hair Wig, one Horse hair wig not
the worse for wearing, one Pale Hair Wig, not worn five times,
marked V. S. E., one brown Natural wig, One old wig of goat’s hair
put in buckle.” Buckle meant to curl; and derivatively a wig was in
buckle when it was rolled on papers for curling. Other
advertisements tell of “Perukes, Tets, and Fox-tails after the
Genteelest Fashion. Ladies’ Tets and wigs in perfect imitation of their
own hair.” Other curious notices are of “Orange Butter” for
“Gentlewomen to comb up their hair with.”
This use of orange butter as a pomatum was certainly unique; it
was really a Dutch marmalade. I read in my “Closet of Rarities,”
dated 1706:—
“The Dutch Way to make Orange-butter. Take new cream
two gallons, beat it up to a thicknesse, then add half a pint of
orange-flower-water, and as much red wine, and so being
become the thicknesse of butter it has both the colour and
smell of an orange.”
A very characteristic and eye-catching advertisement was this
from the “New York Gazette” of May 21, 1750:—
“This is to acquaint the Public, that there is lately arrived
from London the Wonder of the World, an Honest Barber and
Peruke Maker, who might have worked for the King, if his
Majesty would have employed him: It was not for the want of
Money he came here, for he had enough of that at Home, nor
for the want of Business, that he advertises hinself, BUT to
acquaint the Gentlemen and Ladies, That Such a Person is
now in Town, living near Rosemary Lane where Gentlemen
and Ladies may be supplied with Goods as follows, viz.: Tyes,
Full-Bottoms, Majors, Spencers, Fox-Tails, Ramalies, Tacks,
cut and bob Perukes: Also Ladies Tatematongues and Towers
after the Manner that is now wore at Court. By their Humble
and Obedient Servant,
“John Still.”
With the change from simple Dutch ways of hairdressing came in
other details more constrained modes of dressing. With the wig-
maker came the stay-maker, whose curious advertisements may be
read in scores in the provincial newspapers; and his arbitrary
fashions bring us to modern times.
From the deacons’ records of the Dutch Reformed Church at
Albany we catch occasional hints of the dress of the children of the
Dutch colonists. There was no poor-house, and few poor; but since
the church occasionally helped worthy folk who were not rich, we
find the deacons in 1665 and 1666 paying for blue linen for
schorteldoecykers, or aprons, for Albany kindeken; also for haaken
en oogen, or hooks and eyes, for warm under-waists called
borsrockyen. They bought linen for luyers, which were neither
pinning-blankets nor diapers, but a sort of swaddling clothes, which
evidently were worn then by Dutch babies. Voor-schooten, which
were white bibs; neerstucken, which were tuckers, also were worn
by little children. Some little Hans of Pieter had given to him by the
deacons a fine little scarlet aperock, or monkey-jacket; and other
children were furnished linen cosynties, or night-caps with capes.
Yellow stockings were sold at the same time for children, and a gay
little yellow turkey-legged Dutchman in a scarlet monkey-jacket and
fat little breeches must have been a jolly sight.
CHAPTER X
HOLIDAYS
The most important holidays of the early years of the colony were,
apparently, New Year’s Day and May Day, for we find them named
through frequent legislation about rioting on these days, repairing of
damages, etc. It has been said that New Yorkers owe to the Dutch
an everlasting gratitude for our high-stoop houses and the delights of
over two centuries of New Year’s calling. The latter custom lived long
and happily in our midst, died a lingering and lamented death, is still
much honored in our memory, and its extinction deeply deplored and
unwillingly accepted.
The observance of New Year’s Day was, without doubt, followed
by both Dutch and English from the earliest settlement. We know
that Governor Stuyvesant received New Year’s calls, and we also
know that he prohibited excessive “drunken drinking,” unnecessary
firing of guns, and all disorderly behavior on that day. The reign of
the English did not abolish New Year’s visits; and we find Charles
Wolley, an English chaplain, writing in his journal in New York in
1701, of the addition of the English custom of exchange of gifts:—
“The English in New York observed one anniversary custom
and that without superstition, I mean the strenarum
commercium, as Suetonius calls them, a neighborly
commerce of presents every New Year’s Day. Some would
send me a sugar-loaf, some a pair of gloves, some a bottle or
two of wine.”
A further celebration of the day by men in New York was by going
in parties to Beekman’s Swamp to shoot at turkeys.
New Year’s calling was a new fashion to General Washington
when he came to New York to live for a short time, but he adopted it
with approval; and his New Year’s Receptions were imposing
functions.
For a long time the New Year was ushered in, in country towns,
with great noise as well as rejoicing. All through the day groups of
men would go from house to house firing salutes, and gathering
gradually into large parties by recruits from each house until the end
of the day was spent in firing at a mark. The Legislature in March,
1773, attempted to stop the gun-firing, asserting that “great damages
are frequently done on the eve of the last day of December and on
the first and second days of January by persons going from house to
house with guns and other firearms.” In 1785 a similar enactment
was passed by the State Legislature.
In the palmiest days of New Year’s calling, New York City
appeared one great family reunion. Every wheeled vehicle in the
town seemed to be loaded with visitors going from house to house.
Great four and six horse stages packed with hilarious mobs of men
went to the house of every acquaintance of every one in the stage.
Target companies had processions; political bodies called on families
whose head was well known in political life. The newspaper-carriers
brought out addresses yards long with rhymes:—