Scribd Logo
Upload

Welcome to Scribd!

  • SMS Spam Classifier Using Machine Learning

    Uploaded by

    International Journal of Innovative Science and Research Technology
    30 views2 pages
    :- The purpose of this research paper is to examine how machine

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    :- The purpose of this research paper is to examine how machine

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    30 views2 pages

    SMS Spam Classifier Using Machine Learning

    Uploaded by

    International Journal of Innovative Science and Research Technology
    :- The purpose of this research paper is to examine how machine

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Flag for inappropriate content
    of 2

    Volume 8, Issue 6, June – 2023 International Journal of Innovative Science and Research Technology

    ISSN No:-2456-2165

    SMS Spam Classifier Using Machine Learning


    Shaniya Chauhan
    Department of Computer Applications

    Abstract:- The purpose of this research paper is to Till then the effect was not so widespread because
    examine how machine learning techniques are used to internet was only confined to defense setups, large
    identify whether a SMS is spam or not. The growth of international companies and research communities.
    mobile users has led to a dramatic increase in SMS and
    messages. Despite the fact that our idea of information In 1996, when internet was launched for the public, it
    channels are currently seen as spotless and reliable in immediately became popular among the masses and they
    many parts of the world on going data clearly slowly became dependent on it to an extent that it have
    demonstrates that the amount of cell phones Spam is changed their lifestyle. As on April 2018, the number of
    dramatically increasing overtime. It is a growing mobile broadband subscribers in India reached 401.41
    catastrophe, especially in the middle East and Asia. million. Overall, the number of broadband subscribers,
    including wired, in India reached 419.79 million by the end
    Separating SMS spam is a similarly lead task to of April 2018.
    solve this problem. It games several concerns and
    practical fixes from SMS spam separation in any case, it III. METHODOLOGY
    brings up it's own unique problems.
    The terms "Spam" refers to unwanted content with
    Keywords – Machine Learning, Spam Separation. questionable information. The data set consists of a
    randomly choseniction of plain text emails that have been
    I. INTRODUCTION classified as either SPAM or HAM. The model for
    categorising emails as ham and spam is developed using the
    The bulk delivery of unscheduled messages, primarily training data. The model created using the training data is
    of a business nature but also containing offensive content, examined using the test data for accuracy.The system is
    has become a major problem for SMS service for internet shown as a block diagram in Fig.1.
    service providers, businesses,and individual customers in the
    last 10 years due to the spam phenomenon's steady growth.
    Recent analysis show that more than 60% of all messages
    are spam. Spam puts excessive strain on sms frameworks
    ability to make data quickly and store data on servers,
    increasing annual costs for partnership by more than several
    billion dollars.

    II. LITERATURE REVIEW

    The expression 'cybercrime' is a product of the


    expansion in communications technology which has
    accelerated over the last twenty five years.

    Cyber Crime is a term that encompasses a variety of Fig. 1 System Block Diagram
    offences associated with the use of information and
    communication technology. The internet was born around Data cleaning-One of the crucial components of
    1960’s where its access was limited to few scientist, machine learning is data cleaning, it is crucial to the process
    researchers and the defense only. of creating a model. There are no hidden twist or secrets to
    discover, but it is also not the most fancy aspect of machine
    Internet user base have evolved exponentially. Initially learning.
    the computer crime was only confined to making a physical
    damage to the computer and related infrastructure. Exploratory Data Analysis-Data analysis utilizing
    visual methods is called exploratory data analysis. With the
    Around 1980’s the trend changed from causing the use of statistical summaries and graphical representations, it
    physical damaging to computers to making a computer is used to identify trends, patterns, or to verify assumptions.
    malfunction using a malicious code called virus.

    IJISRT23JUN164 www.ijisrt.com 29
    Volume 8, Issue 6, June – 2023 International Journal of Innovative Science and Research Technology
    ISSN No:-2456-2165
    Data pre-processing-The learning data's emails are in and rearranging the formula for determining the interesting
    plain text format. The simple text must be transformed into rate can be made at a later time.
    characteristics that can represent emails. We may then apply
    a learning algorithm to the emails using these features. First We intend to tackle more difficult issue in the future,
    and for most certain preposition procedures are carried out. like the analysis and administration of report data stored in
    Model building-The last step is model building where we SMS spam filters.Future work will additionally concentrate
    finalize everything and build a precise model. on finding a solution to this issue.

    IV. DATASET AND FEATURES REFERENCES

    The dataset used was picked up from Kaggle. This [1]. https://www.youtube.com/watch?v=YncZ0WwxyzU
    dataset was first thoroughly examined and then data [2]. Swayam
    cleaning and data analysis was performed. Further [3]. 3.https://www.researchgate.net/publication/349799157
    Exploratory data analysis was done for this dataset. The data _SMS_Spam_Detection_Using_Machine_Learning
    pre-processing includes the removal of stop words, [4]. https://ieeexplore.ieee.org/document/9441783
    stemming, tokenization and removal of special characters
    and punctuation marks.

    V. RESULT

    We tested different algorithms for the dataset for the


    best precision score and accuracy.

    Typically, the performance of an SMS spam classifier


    is evaluated using metrics such as precision, recall, and F1
    score. Precision is the fraction of spam messages correctly
    identified as spam, while recall is the fraction of all spam
    messages that are correctly identified by the classifier. The
    F1 score is the harmonic mean of precision and recall, which
    provides an overall measure of the classifier's performance.

    Multinomial Naive Bayes is a popular machine


    learning algorithm used for text classification tasks,
    including SMS spam classification. The algorithm works by
    modeling the probability distribution of words in a text
    message and calculating the likelihood that a message is
    spam or not based on these probabilities.

    The performance of a Multinomial Naive Bayes SMS


    spam classifier can be evaluated using metrics such as
    accuracy, precision, recall, and F1 score. The highest
    precision achieved is 1.0 and accuracy is about 97 percent.

    VI. CONCLUSION & FUTURE WORK

    This projects goal and objectives were established at


    the very very of the process and were accomplished
    throughout . The research process involves a detailed
    analysis of the various filtering algorithms and available
    anti-spam technologies in order to gather all the
    information.

    The projects work was inspired in part by the large


    scale research articles and existing software packages
    mentioned above. The entire project was broken up into
    various iterations.Each iteration was finished by walking
    through the different phases.

    There are still certain areas that can be improved, such


    as by incorporating more filtering methods or altering
    certain features of the ones that already exist. Changes like
    increasing or decreasing the message's intriguing word count

    IJISRT23JUN164 www.ijisrt.com 30

    You might also like