TY - JOUR AU - Puspita, Ari AU - Jefi, Jefi AU - Fahmi, Muhammad PY - 2019/02/13 Y2 - 2024/03/28 TI - Prediksi Peminatan Pelanggan dalam Penjualan Produk Sepatu Menggunakan Metode Decision Tree Berbasis Particle Swarm Optimization pada PT. Baskara Cipta Pratama JF - Jurnal Teknik Informatika JA - JTI VL - 5 IS - 1 SE - DO - 10.51998/jti.v5i1.296 UR - https://ejournal.antarbangsa.ac.id/jti/article/view/296 SP - 10-17 AB - <p><strong><em>Abstract</em></strong><strong>—</strong><strong><em>T</em></strong><strong><em>he PSO-based optimization C4.5 model gives a higher value of 78.16% compared to the C4.5 algorithm model that is 73.88</em></strong><strong><em>. The </em></strong><strong><em>results obtained difference</em></strong><strong><em>s</em></strong><strong><em> between the two models </em></strong><strong><em>by</em></strong><strong><em> 4.28%. While for evaluation using ROC curve for second model that is, for model of algorithm C4.5 value of AUC is 0,764 with level of diagnosis classification fair, and for model of algorithm C4.5 based on PSO </em></strong><strong><em>A</em></strong><strong><em>UC is 0,780 with level of diagnosis of fair classification. </em></strong><strong><em>It s concluded that</em></strong><strong><em> ROC curvesmodels</em></strong><strong><em> shows</em></strong><strong><em> C4.5</em></strong><strong><em> algorithm</em></strong><strong><em> based on PSO </em></strong><strong><em>is larger. It </em></strong><strong><em>can be inferred</em></strong><strong><em> that</em></strong><strong><em> </em></strong><strong><em>C4.5</em></strong><strong><em> algorithm </em></strong><strong><em>based</em></strong><strong><em> on</em></strong><strong><em> particle</em></strong><strong><em> swam</em></strong><strong><em> optimization is more accurate in predicting the </em></strong><strong><em>customers’</em></strong><strong><em> interest </em></strong><strong><em>for buying shoes</em></strong><strong><em>.</em></strong><strong></strong></p><p><strong><em> </em></strong></p><p class="NoSpacing1"><strong>        </strong><strong>I</strong><strong>ntisari</strong><strong>—</strong> <strong>A</strong><strong>nalisis optimasi model algoritma C4.5 berbasis PSO memberikan nilai akurasi yang lebih tinggi yaitu 78.16% dibandingkan dengan model algoritma C4.5 yaitu 73.88%. Dari hasil tersebut didapatkan selisih antara kedua model yaitu 4,28%. Sementara untuk evalusai menggunakan ROC <em>curve</em> untuk kedua model yaitu, untuk model algoritma C4.5 nilai AUC adalah 0.764 dengan tingkat diagnosa </strong><strong><em>Fair classification</em></strong><strong>, dan untuk model algoritma C4.5 berbasis PSO nilai AUC adalah 0.780 dengan tingkat diagnosa </strong><strong><em>Fair classification. </em></strong><strong>Dari evaluasi ROC <em>curve</em> tersebut terlihat bahwa model algoritma C4.5 berbasis PSO lebih besar  Sehingga dapat disimpulkan bahwa algoritma C4.5 berbasis <em>particle swarm optimization</em> lebih akurat dalam memprediksi </strong><strong>minat beli prod</strong><strong>u</strong><strong>k sepatu.</strong></p><p><strong> </strong></p><p> </p><strong>Kata Kunci<em> </em></strong>—<strong><em> </em>C4.5, Produk, Sepatu PSO </strong> ER -