ChristopherEban/Prediction-of-Legal-Decisions-Using-ML-Dataset-of-Nigeria-Appeal-Court-Case-Authorities-

Application of Machine Learning to Predict Legal Outcomes in Uncontested/Contested Court Cases Dataset Description: The dataset relates to legal decisions of Nigerian Courts, consisting of 500 rows and 6 columns. The columns include CASE TITLE, SUIT NO, CITATION, STORY, ISSUES, and an S/N column mostly populated with NaN values. • CASE TITLE: Contains names of parties involved in each case (formatted as "Plaintiff v. Defendant" or "Appellant v. Respondent"). It features 480 unique entries with no missing values. • SUIT NO: Serves as a unique identifier for each case, with all 500 entries being unique and free of missing values. • CITATION: Provides specific references for each case, generally formatted as "(Year) LPELR-XXXXX(CA)." All 500 citations are unique and contain no missing values. • STORY: Includes detailed narratives or introductions and Facts about each case, providing background information and context. Issues and Decisions held by the Appeal court for each of the cases. There are 499 unique stories with one missing value. • ISSUES: Lists the key legal issues or points of law addressed in each case, with 498 unique entries and no missing values. The attributes and their data types indicate that most data is textual, suitable for Natural Language Processing (NLP) applications to extract insights and predict legal outcomes. The dataset is well-organized and detailed, making it suitable for legal research, predictive modeling, and text analysis. This dataset is sourced from a pool originally hosted by Law Pavillion Electronic Law Report, a Nigerian-registered and operating online law reporting company. Link: docs.google.com Prediction of Legal Decisions using Machine Learning (1).xlsx Steps: Data Loading and Initial Exploration: • The code begins by importing necessary libraries (pandas, re, nltk, stopwords, WordNetLemmatizer, TfidfVectorizer, matplotlib.pyplot, seaborn, WordCloud, tensorflow, Sequential, Dense, Dropout) for data handling, text preprocessing, vi …