<![CDATA[The mathematics learning outcomes of fifth-grade students at SDN 14 Gadut, Agam Regency, on the topic of data representation were still unsatisfactory. Only 9 out of 16 students (56,25%) achieved the Minimum Learning Achievement Criteria (KKTP) of 75. This condition indicated the need for a learning model that actively engages students and promotes contextual problem-solving. This study aimed to improve the mathematics learning outcomes of fifth-grade students on the topic of data representation through the implementation of the Problem Based Learning (PBL) model integrated with the Deep Learning approach. This research employed a Classroom Action Research (CAR) design based on the Kemmis and McTaggart model, conducted in two cycles. The participants consisted of all 16 fifth-grade students, including 10 girls and 6 boys. Data were collected through students' learning outcomes from worksheets (LKPD) and achievement tests, periodic practice exercises, as well as attitude and skills assessment questionnaires. Quantitative data were analyzed using mean scores and learning mastery percentages, while qualitative data were examined through data reduction, data presentation, and conclusion drawing. The findings revealed that learning mastery improved from 50% at the initial stage to 93.75% in Cycle I Meeting I and remained relatively stable at 87.50% until the end of Cycle II. The average cognitive achievement increased from 84.42 to 88.99 during Cycle I but slightly declined to 84.23 in Cycle II. In contrast, students' skills demonstrated consistent improvement, increasing from 71.88 to 76.33. Overall, the results indicate that integrating the Problem Based Learning (PBL) model with the Deep Learning approach contributed positively to improving students' mathematics learning outcomes on the topic of data representation, although the improvement in cognitive achievement was not consistently sustained across all cycles. ABSTRAK Hasil belajar matematika siswa kelas V SDN 14 Gadut Kabupaten Agam pada topik penyajian data masih rendah, terbukti hanya 9 dari 16 siswa (56,25%), yang mencapai Kriteria Ketercapaian Tujuan Pembelajaran (KKTP) sebesar 75, sehingga dibutuhkan model pembelajaran yang mendorong keterlibatan aktif dan penyelesaian masalah kontekstual. Penelitian ini bertujuan untuk menghasilkan peningkatan hasil belajar matematika murid kelas V SDN 14 Gadut. pada topik penyajian data melalui penggunaan model Problem Based Learning (PBL) yang diperkuat pendekatan Deep Learning. Penelitian Tindakan Kelas (PTK) berdasarkan model Kemmis dan McTaggart, dilakukan dalam dua siklus, dengan subjek seluruh murid kelas V berjumlah 16 orang (10 perempuan, 6 laki-laki). Data diperoleh melalui hasil belajar (LKPD dan evaluasi), latihan soal berkala, serta angket sikap dan keterampilan, dianalisis secara kuantitatif berupa persentase ketuntasan dan rata-rata serta kualitatif berdasarkan pengelompokan data, pemaparan data, dan penyimpulan. Berdasarkan hasil penelitian menunjukkan ketuntasan belajar meningkat dari 50% dari tahap awal menjadi 93,75% pada siklus I pertemuan I, kemudian relative stabil pada 87,5% hinga siklus II. Rata-rata pengetahuan meningkat dari 84,42 menjadi 88,99 selama siklus I, namun menurun menjadi 84,23 pada siklus II, sementara keterampilan meningkat konsisten dari 71,88 menjadi 76. Simpulan menunjukkan bahwa implementasi PBL terintegrasi Deep Learning berkontribusi dalam meningkatkan capaian belajar matematika siswa pada topik penyajian data, meskipun capaian pengetahuan belum konsisten meningkat pada seluruh siklus.]]>
