Predicting Classes of helpdesk service ticket priority Based on Support Vector Machine

Project info

• Field of studyInformatics and Information Systems
• LevelMaster's thesis (S2)
• Method typeClassification
• Estimated build time± 5 working days
• Toolsscikit-learn SVC, GridSearchCV, StandardScaler, DecisionBoundaryDisplay
• PriceUS$25 - US$156

Research goal

Classify ticket priority from resolution time and escalation count with Support Vector Machine, then assess performance using accuracy and the confusion matrix.

Variables

Example result

• 0.82Accuracy
• 0.786Precision
• 0.88Recall
• 0.83F1-score
• 0.899ROC AUC
• 51Support vectors

The RBF-kernel SVM model classifies Ticket priority (low/high) from Resolution time and Escalation count with test accuracy 0.820, F1 0.830 and AUC 0.899. It uses 51 support vectors.

Computed from illustrative synthetic data, not from a real study.

About Support Vector Machine

A Support Vector Machine finds the separating hyperplane with the largest margin, and kernels allow nonlinear decision boundaries. It suits classification with small to medium data and high-dimensional features.

Steps

1. 1Standardise features
2. 2Split into stratified training and test sets
3. 3Tune C and gamma (RBF kernel) with GridSearchCV
4. 4Train SVC with the best parameters
5. 5Evaluate accuracy, F1, and the confusion matrix
6. 6Plot the decision boundary if there are two features

Assumptions to check

• Features must be standardised because SVM is scale sensitive
• C and gamma must be tuned by cross-validation
• SVM probability outputs are only calibrated estimates

Reading and reporting the results

Report the kernel, best parameters, test metrics, and the number of support vectors. Explain that the model is a black box, so feature interpretation is limited.

Example Python code

A short starting script with this project's variable names. The delivered version is adapted to your data and fully commented.

import numpy as np
import matplotlib.pyplot as plt
from sklearn.svm import SVC
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import make_pipeline
from sklearn.model_selection import GridSearchCV, train_test_split
from sklearn.metrics import accuracy_score, f1_score

rng = np.random.default_rng(41)
n = 300
resolution_time = rng.normal(0, 1, n)
escalation_count = rng.normal(0, 1, n)
ticket_priority = ((resolution_time ** 2 + escalation_count ** 2) > 1.2).astype(int)

X = np.column_stack([resolution_time, escalation_count])
X_tr, X_te, y_tr, y_te = train_test_split(X, ticket_priority, test_size=0.25, stratify=ticket_priority, random_state=41)
pipe = make_pipeline(StandardScaler(), SVC(kernel="rbf"))
grid = {"svc__C": [0.1, 1, 10, 100], "svc__gamma": [0.01, 0.1, 1]}
gs = GridSearchCV(pipe, grid, cv=5).fit(X_tr, y_tr)
pred = gs.predict(X_te)
print("Best params:", gs.best_params_)
print("Accuracy:", round(accuracy_score(y_te, pred), 3), "F1:", round(f1_score(y_te, pred), 3))
print("Support vectors:", gs.best_estimator_[-1].n_support_)

gx, gy = np.meshgrid(np.linspace(-3, 3, 200), np.linspace(-3, 3, 200))
zz = gs.predict(np.c_[gx.ravel(), gy.ravel()]).reshape(gx.shape)
plt.contourf(gx, gy, zz, alpha=0.3)
plt.scatter(X_te[:, 0], X_te[:, 1], c=y_te, edgecolor="k")
plt.xlabel("x1"); plt.ylabel("x2")
plt.show()

Packages and prices

Choose how much help you need. Each higher package includes everything in the lower ones. Prices exclude 11% VAT.

• Payment: 50% down payment before work starts, balance after the deliverables are handed over
• Tax: Prices exclude 11% VAT
• Data and academic integrity: Results are computed from your own research data; you write and stand behind your manuscript, we help with the analysis and your understanding of it
• Revisions: Up to 2 revisions are included in the price (Package 3: 3 rounds)

Frequently asked questions